docs: Add Quick Start section (issue #13)

Bumps the actions group with 2 updates: [actions/checkout](https://github.com/actions/checkout) and [softprops/action-gh-release](https://github.com/softprops/action-gh-release).


Updates `actions/checkout` from 4 to 7
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](https://github.com/actions/checkout/compare/v4...v7)

Updates `softprops/action-gh-release` from 2 to 3
- [Release notes](https://github.com/softprops/action-gh-release/releases)
- [Changelog](https://github.com/softprops/action-gh-release/blob/master/CHANGELOG.md)
- [Commits](https://github.com/softprops/action-gh-release/compare/v2...v3)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-version: '7'
  dependency-type: direct:production
  update-type: version-update:semver-major
  dependency-group: actions
- dependency-name: softprops/action-gh-release
  dependency-version: '3'
  dependency-type: direct:production
  update-type: version-update:semver-major
  dependency-group: actions
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

.github/dependabot.yml

@@ -0,0 +1,15 @@
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"
+    commit-message:
+      prefix: "chore"
+    labels:
+      - "dependencies"
+      - "github-actions"
+    groups:
+      actions:
+        patterns:
+          - "*"

.github/workflows/hassfest.yaml

@@ -26,19 +26,15 @@ jobs:
     timeout-minutes: 10
 
     steps:
-      - name: ⤵️ Check out code from GitHub
-        uses: actions/checkout@v4
+      - name: Check out code from GitHub
+        uses: actions/checkout@v7
         with:
           fetch-depth: 0
+          persist-credentials: false
 
-      - name: 🚀 Run hassfest validation
+      - name: Run hassfest validation
         uses: home-assistant/actions/hassfest@master
 
-      - name: ℹ️ Print hassfest version
-        if: always()
-        run: |
-          echo "Hassfest version: $(hassfest --version)"
-
     concurrency:
       group: ${{ github.workflow }}-${{ github.ref }}
       cancel-in-progress: true

.github/workflows/release.yml

@@ -0,0 +1,38 @@
+name: Release
+
+on:
+  release:
+    types: [created]
+
+permissions:
+  contents: write
+
+jobs:
+  build:
+    name: Build and upload release asset
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v7
+        with:
+          ref: ${{ github.event.release.tag_name }}
+          persist-credentials: false
+
+      - name: Create zip archive
+        run: |
+          cd custom_components
+          zip -r ../ha_text_ai.zip ha_text_ai \
+            -x "ha_text_ai/__pycache__/*" \
+            -x "*.pyc" \
+            -x "*.pyo" \
+            -x "*/__pycache__/*" \
+            -x "*.DS_Store"
+
+      - name: Upload release asset
+        uses: softprops/action-gh-release@v3
+        with:
+          tag_name: ${{ github.event.release.tag_name }}
+          files: ha_text_ai.zip
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -0,0 +1,40 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+dist/
+build/
+*.egg
+
+# Virtual environments
+.venv/
+venv/
+
+# IDE
+.idea/
+.vscode/
+*.swp
+*.swo
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Archives
+*.zip
+
+# Home Assistant
+.storage/
+
+# Claude Code working files
+handoff.*.md
+CLAUDE.md
+AGENTS.md
+GEMINI.md
+.claude/
+.cursor/
+.cursorrules
+.windsurfrules
+docs/specs/
+docs/superpowers/
+docs/plans/

LICENSE

@@ -1,437 +1,21 @@
-Attribution-NonCommercial-ShareAlike 4.0 International
-
-=======================================================================
-
-Creative Commons Corporation ("Creative Commons") is not a law firm and
-does not provide legal services or legal advice. Distribution of
-Creative Commons public licenses does not create a lawyer-client or
-other relationship. Creative Commons makes its licenses and related
-information available on an "as-is" basis. Creative Commons gives no
-warranties regarding its licenses, any material licensed under their
-terms and conditions, or any related information. Creative Commons
-disclaims all liability for damages resulting from their use to the
-fullest extent possible.
-
-Using Creative Commons Public Licenses
-
-Creative Commons public licenses provide a standard set of terms and
-conditions that creators and other rights holders may use to share
-original works of authorship and other material subject to copyright
-and certain other rights specified in the public license below. The
-following considerations are for informational purposes only, are not
-exhaustive, and do not form part of our licenses.
-
-     Considerations for licensors: Our public licenses are
-     intended for use by those authorized to give the public
-     permission to use material in ways otherwise restricted by
-     copyright and certain other rights. Our licenses are
-     irrevocable. Licensors should read and understand the terms
-     and conditions of the license they choose before applying it.
-     Licensors should also secure all rights necessary before
-     applying our licenses so that the public can reuse the
-     material as expected. Licensors should clearly mark any
-     material not subject to the license. This includes other CC-
-     licensed material, or material used under an exception or
-     limitation to copyright. More considerations for licensors:
-    wiki.creativecommons.org/Considerations_for_licensors
-
-     Considerations for the public: By using one of our public
-     licenses, a licensor grants the public permission to use the
-     licensed material under specified terms and conditions. If
-     the licensor's permission is not necessary for any reason--for
-     example, because of any applicable exception or limitation to
-     copyright--then that use is not regulated by the license. Our
-     licenses grant only permissions under copyright and certain
-     other rights that a licensor has authority to grant. Use of
-     the licensed material may still be restricted for other
-     reasons, including because others have copyright or other
-     rights in the material. A licensor may make special requests,
-     such as asking that all changes be marked or described.
-     Although not required by our licenses, you are encouraged to
-     respect those requests where reasonable. More considerations
-     for the public:
-    wiki.creativecommons.org/Considerations_for_licensees
-
-=======================================================================
-
-Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
-Public License
-
-By exercising the Licensed Rights (defined below), You accept and agree
-to be bound by the terms and conditions of this Creative Commons
-Attribution-NonCommercial-ShareAlike 4.0 International Public License
-("Public License"). To the extent this Public License may be
-interpreted as a contract, You are granted the Licensed Rights in
-consideration of Your acceptance of these terms and conditions, and the
-Licensor grants You such rights in consideration of benefits the
-Licensor receives from making the Licensed Material available under
-these terms and conditions.
-
-
-Section 1 -- Definitions.
-
-  a. Adapted Material means material subject to Copyright and Similar
-     Rights that is derived from or based upon the Licensed Material
-     and in which the Licensed Material is translated, altered,
-     arranged, transformed, or otherwise modified in a manner requiring
-     permission under the Copyright and Similar Rights held by the
-     Licensor. For purposes of this Public License, where the Licensed
-     Material is a musical work, performance, or sound recording,
-     Adapted Material is always produced where the Licensed Material is
-     synched in timed relation with a moving image.
-
-  b. Adapter's License means the license You apply to Your Copyright
-     and Similar Rights in Your contributions to Adapted Material in
-     accordance with the terms and conditions of this Public License.
-
-  c. BY-NC-SA Compatible License means a license listed at
-     creativecommons.org/compatiblelicenses, approved by Creative
-     Commons as essentially the equivalent of this Public License.
-
-  d. Copyright and Similar Rights means copyright and/or similar rights
-     closely related to copyright including, without limitation,
-     performance, broadcast, sound recording, and Sui Generis Database
-     Rights, without regard to how the rights are labeled or
-     categorized. For purposes of this Public License, the rights
-     specified in Section 2(b)(1)-(2) are not Copyright and Similar
-     Rights.
-
-  e. Effective Technological Measures means those measures that, in the
-     absence of proper authority, may not be circumvented under laws
-     fulfilling obligations under Article 11 of the WIPO Copyright
-     Treaty adopted on December 20, 1996, and/or similar international
-     agreements.
-
-  f. Exceptions and Limitations means fair use, fair dealing, and/or
-     any other exception or limitation to Copyright and Similar Rights
-     that applies to Your use of the Licensed Material.
-
-  g. License Elements means the license attributes listed in the name
-     of a Creative Commons Public License. The License Elements of this
-     Public License are Attribution, NonCommercial, and ShareAlike.
-
-  h. Licensed Material means the artistic or literary work, database,
-     or other material to which the Licensor applied this Public
-     License.
-
-  i. Licensed Rights means the rights granted to You subject to the
-     terms and conditions of this Public License, which are limited to
-     all Copyright and Similar Rights that apply to Your use of the
-     Licensed Material and that the Licensor has authority to license.
-
-  j. Licensor means the individual(s) or entity(ies) granting rights
-     under this Public License.
-
-  k. NonCommercial means not primarily intended for or directed towards
-     commercial advantage or monetary compensation. For purposes of
-     this Public License, the exchange of the Licensed Material for
-     other material subject to Copyright and Similar Rights by digital
-     file-sharing or similar means is NonCommercial provided there is
-     no payment of monetary compensation in connection with the
-     exchange.
-
-  l. Share means to provide material to the public by any means or
-     process that requires permission under the Licensed Rights, such
-     as reproduction, public display, public performance, distribution,
-     dissemination, communication, or importation, and to make material
-     available to the public including in ways that members of the
-     public may access the material from a place and at a time
-     individually chosen by them.
-
-  m. Sui Generis Database Rights means rights other than copyright
-     resulting from Directive 96/9/EC of the European Parliament and of
-     the Council of 11 March 1996 on the legal protection of databases,
-     as amended and/or succeeded, as well as other essentially
-     equivalent rights anywhere in the world.
-
-  n. You means the individual or entity exercising the Licensed Rights
-     under this Public License. Your has a corresponding meaning.
-
-
-Section 2 -- Scope.
-
-  a. License grant.
-
-       1. Subject to the terms and conditions of this Public License,
-          the Licensor hereby grants You a worldwide, royalty-free,
-          non-sublicensable, non-exclusive, irrevocable license to
-          exercise the Licensed Rights in the Licensed Material to:
-
-            a. reproduce and Share the Licensed Material, in whole or
-               in part, for NonCommercial purposes only; and
-
-            b. produce, reproduce, and Share Adapted Material for
-               NonCommercial purposes only.
-
-       2. Exceptions and Limitations. For the avoidance of doubt, where
-          Exceptions and Limitations apply to Your use, this Public
-          License does not apply, and You do not need to comply with
-          its terms and conditions.
-
-       3. Term. The term of this Public License is specified in Section
-          6(a).
-
-       4. Media and formats; technical modifications allowed. The
-          Licensor authorizes You to exercise the Licensed Rights in
-          all media and formats whether now known or hereafter created,
-          and to make technical modifications necessary to do so. The
-          Licensor waives and/or agrees not to assert any right or
-          authority to forbid You from making technical modifications
-          necessary to exercise the Licensed Rights, including
-          technical modifications necessary to circumvent Effective
-          Technological Measures. For purposes of this Public License,
-          simply making modifications authorized by this Section 2(a)
-          (4) never produces Adapted Material.
-
-       5. Downstream recipients.
-
-            a. Offer from the Licensor -- Licensed Material. Every
-               recipient of the Licensed Material automatically
-               receives an offer from the Licensor to exercise the
-               Licensed Rights under the terms and conditions of this
-               Public License.
-
-            b. Additional offer from the Licensor -- Adapted Material.
-               Every recipient of Adapted Material from You
-               automatically receives an offer from the Licensor to
-               exercise the Licensed Rights in the Adapted Material
-               under the conditions of the Adapter's License You apply.
-
-            c. No downstream restrictions. You may not offer or impose
-               any additional or different terms or conditions on, or
-               apply any Effective Technological Measures to, the
-               Licensed Material if doing so restricts exercise of the
-               Licensed Rights by any recipient of the Licensed
-               Material.
-
-       6. No endorsement. Nothing in this Public License constitutes or
-          may be construed as permission to assert or imply that You
-          are, or that Your use of the Licensed Material is, connected
-          with, or sponsored, endorsed, or granted official status by,
-          the Licensor or others designated to receive attribution as
-          provided in Section 3(a)(1)(A)(i).
-
-  b. Other rights.
-
-       1. Moral rights, such as the right of integrity, are not
-          licensed under this Public License, nor are publicity,
-          privacy, and/or other similar personality rights; however, to
-          the extent possible, the Licensor waives and/or agrees not to
-          assert any such rights held by the Licensor to the limited
-          extent necessary to allow You to exercise the Licensed
-          Rights, but not otherwise.
-
-       2. Patent and trademark rights are not licensed under this
-          Public License.
-
-       3. To the extent possible, the Licensor waives any right to
-          collect royalties from You for the exercise of the Licensed
-          Rights, whether directly or through a collecting society
-          under any voluntary or waivable statutory or compulsory
-          licensing scheme. In all other cases the Licensor expressly
-          reserves any right to collect such royalties, including when
-          the Licensed Material is used other than for NonCommercial
-          purposes.
-
-
-Section 3 -- License Conditions.
-
-Your exercise of the Licensed Rights is expressly made subject to the
-following conditions.
-
-  a. Attribution.
-
-       1. If You Share the Licensed Material (including in modified
-          form), You must:
-
-            a. retain the following if it is supplied by the Licensor
-               with the Licensed Material:
-
-                 i. identification of the creator(s) of the Licensed
-                    Material and any others designated to receive
-                    attribution, in any reasonable manner requested by
-                    the Licensor (including by pseudonym if
-                    designated);
-
-                ii. a copyright notice;
-
-               iii. a notice that refers to this Public License;
-
-                iv. a notice that refers to the disclaimer of
-                    warranties;
-
-                 v. a URI or hyperlink to the Licensed Material to the
-                    extent reasonably practicable;
-
-            b. indicate if You modified the Licensed Material and
-               retain an indication of any previous modifications; and
-
-            c. indicate the Licensed Material is licensed under this
-               Public License, and include the text of, or the URI or
-               hyperlink to, this Public License.
-
-       2. You may satisfy the conditions in Section 3(a)(1) in any
-          reasonable manner based on the medium, means, and context in
-          which You Share the Licensed Material. For example, it may be
-          reasonable to satisfy the conditions by providing a URI or
-          hyperlink to a resource that includes the required
-          information.
-       3. If requested by the Licensor, You must remove any of the
-          information required by Section 3(a)(1)(A) to the extent
-          reasonably practicable.
-
-  b. ShareAlike.
-
-     In addition to the conditions in Section 3(a), if You Share
-     Adapted Material You produce, the following conditions also apply.
-
-       1. The Adapter's License You apply must be a Creative Commons
-          license with the same License Elements, this version or
-          later, or a BY-NC-SA Compatible License.
-
-       2. You must include the text of, or the URI or hyperlink to, the
-          Adapter's License You apply. You may satisfy this condition
-          in any reasonable manner based on the medium, means, and
-          context in which You Share Adapted Material.
-
-       3. You may not offer or impose any additional or different terms
-          or conditions on, or apply any Effective Technological
-          Measures to, Adapted Material that restrict exercise of the
-          rights granted under the Adapter's License You apply.
-
-
-Section 4 -- Sui Generis Database Rights.
-
-Where the Licensed Rights include Sui Generis Database Rights that
-apply to Your use of the Licensed Material:
-
-  a. for the avoidance of doubt, Section 2(a)(1) grants You the right
-     to extract, reuse, reproduce, and Share all or a substantial
-     portion of the contents of the database for NonCommercial purposes
-     only;
-
-  b. if You include all or a substantial portion of the database
-     contents in a database in which You have Sui Generis Database
-     Rights, then the database in which You have Sui Generis Database
-     Rights (but not its individual contents) is Adapted Material,
-     including for purposes of Section 3(b); and
-
-  c. You must comply with the conditions in Section 3(a) if You Share
-     all or a substantial portion of the contents of the database.
-
-For the avoidance of doubt, this Section 4 supplements and does not
-replace Your obligations under this Public License where the Licensed
-Rights include other Copyright and Similar Rights.
-
-
-Section 5 -- Disclaimer of Warranties and Limitation of Liability.
-
-  a. UNLESS OTHERWISE SEPARATELY UNDERTAKEN BY THE LICENSOR, TO THE
-     EXTENT POSSIBLE, THE LICENSOR OFFERS THE LICENSED MATERIAL AS-IS
-     AND AS-AVAILABLE, AND MAKES NO REPRESENTATIONS OR WARRANTIES OF
-     ANY KIND CONCERNING THE LICENSED MATERIAL, WHETHER EXPRESS,
-     IMPLIED, STATUTORY, OR OTHER. THIS INCLUDES, WITHOUT LIMITATION,
-     WARRANTIES OF TITLE, MERCHANTABILITY, FITNESS FOR A PARTICULAR
-     PURPOSE, NON-INFRINGEMENT, ABSENCE OF LATENT OR OTHER DEFECTS,
-     ACCURACY, OR THE PRESENCE OR ABSENCE OF ERRORS, WHETHER OR NOT
-     KNOWN OR DISCOVERABLE. WHERE DISCLAIMERS OF WARRANTIES ARE NOT
-     ALLOWED IN FULL OR IN PART, THIS DISCLAIMER MAY NOT APPLY TO YOU.
-
-  b. TO THE EXTENT POSSIBLE, IN NO EVENT WILL THE LICENSOR BE LIABLE
-     TO YOU ON ANY LEGAL THEORY (INCLUDING, WITHOUT LIMITATION,
-     NEGLIGENCE) OR OTHERWISE FOR ANY DIRECT, SPECIAL, INDIRECT,
-     INCIDENTAL, CONSEQUENTIAL, PUNITIVE, EXEMPLARY, OR OTHER LOSSES,
-     COSTS, EXPENSES, OR DAMAGES ARISING OUT OF THIS PUBLIC LICENSE OR
-     USE OF THE LICENSED MATERIAL, EVEN IF THE LICENSOR HAS BEEN
-     ADVISED OF THE POSSIBILITY OF SUCH LOSSES, COSTS, EXPENSES, OR
-     DAMAGES. WHERE A LIMITATION OF LIABILITY IS NOT ALLOWED IN FULL OR
-     IN PART, THIS LIMITATION MAY NOT APPLY TO YOU.
-
-  c. The disclaimer of warranties and limitation of liability provided
-     above shall be interpreted in a manner that, to the extent
-     possible, most closely approximates an absolute disclaimer and
-     waiver of all liability.
-
-
-Section 6 -- Term and Termination.
-
-  a. This Public License applies for the term of the Copyright and
-     Similar Rights licensed here. However, if You fail to comply with
-     this Public License, then Your rights under this Public License
-     terminate automatically.
-
-  b. Where Your right to use the Licensed Material has terminated under
-     Section 6(a), it reinstates:
-
-       1. automatically as of the date the violation is cured, provided
-          it is cured within 30 days of Your discovery of the
-          violation; or
-
-       2. upon express reinstatement by the Licensor.
-
-     For the avoidance of doubt, this Section 6(b) does not affect any
-     right the Licensor may have to seek remedies for Your violations
-     of this Public License.
-
-  c. For the avoidance of doubt, the Licensor may also offer the
-     Licensed Material under separate terms or conditions or stop
-     distributing the Licensed Material at any time; however, doing so
-     will not terminate this Public License.
-
-  d. Sections 1, 5, 6, 7, and 8 survive termination of this Public
-     License.
-
-
-Section 7 -- Other Terms and Conditions.
-
-  a. The Licensor shall not be bound by any additional or different
-     terms or conditions communicated by You unless expressly agreed.
-
-  b. Any arrangements, understandings, or agreements regarding the
-     Licensed Material not stated herein are separate from and
-     independent of the terms and conditions of this Public License.
-
-
-Section 8 -- Interpretation.
-
-  a. For the avoidance of doubt, this Public License does not, and
-     shall not be interpreted to, reduce, limit, restrict, or impose
-     conditions on any use of the Licensed Material that could lawfully
-     be made without permission under this Public License.
-
-  b. To the extent possible, if any provision of this Public License is
-     deemed unenforceable, it shall be automatically reformed to the
-     minimum extent necessary to make it enforceable. If the provision
-     cannot be reformed, it shall be severed from this Public License
-     without affecting the enforceability of the remaining terms and
-     conditions.
-
-  c. No term or condition of this Public License will be waived and no
-     failure to comply consented to unless expressly agreed to by the
-     Licensor.
-
-  d. Nothing in this Public License constitutes or may be interpreted
-     as a limitation upon, or waiver of, any privileges and immunities
-     that apply to the Licensor or You, including from the legal
-     processes of any jurisdiction or authority.
-
-=======================================================================
-
-Creative Commons is not a party to its public
-licenses. Notwithstanding, Creative Commons may elect to apply one of
-its public licenses to material it publishes and in those instances
-will be considered the “Licensor.” The text of the Creative Commons
-public licenses is dedicated to the public domain under the CC0 Public
-Domain Dedication. Except for the limited purpose of indicating that
-material is shared under a Creative Commons public license or as
-otherwise permitted by the Creative Commons policies published at
-creativecommons.org/policies, Creative Commons does not authorize the
-use of the trademark "Creative Commons" or any other trademark or logo
-of Creative Commons without its prior written consent including,
-without limitation, in connection with any unauthorized modifications
-to any of its public licenses or any other arrangements,
-understandings, or agreements concerning use of licensed material. For
-the avoidance of doubt, this paragraph does not form part of the
-public licenses.
-
-Creative Commons may be contacted at creativecommons.org.
+MIT License
+
+Copyright (c) 2024-2026 SMKRV
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,182 +1,118 @@
-# 🤖 HA Text AI for Home Assistant
+# HA Text AI for Home Assistant
 
 <div align="center">  
 
-  ![GitHub release](https://img.shields.io/github/v/release/smkrv/ha-text-ai?style=flat-square) ![GitHub last commit](https://img.shields.io/github/last-commit/smkrv/ha-text-ai?style=flat-square) [![License: CC BY-NC-SA 4.0](https://img.shields.io/badge/License-CC%20BY--NC--SA%204.0-lightgrey.svg?style=flat-square)](https://creativecommons.org/licenses/by-nc-sa/4.0/) [![hacs_badge](https://img.shields.io/badge/HACS-Default-41BDF5.svg?style=flat-square)](https://github.com/hacs/integration)
+  ![GitHub release](https://img.shields.io/github/v/release/smkrv/ha-text-ai?style=flat-square) ![GitHub last commit](https://img.shields.io/github/last-commit/smkrv/ha-text-ai?style=flat-square) [![License: MIT](https://img.shields.io/badge/License-MIT-green.svg?style=flat-square)](https://opensource.org/licenses/MIT) [![hacs_badge](https://img.shields.io/badge/HACS-Default-41BDF5.svg?style=flat-square)](https://github.com/hacs/integration)
   ![Deutsch](https://img.shields.io/badge/lang-DE-blue?style=flat-square) ![English](https://img.shields.io/badge/lang-EN-blue?style=flat-square) ![Español](https://img.shields.io/badge/lang-ES-blue?style=flat-square) ![हिन्दी](https://img.shields.io/badge/lang-HI-blue?style=flat-square) ![Italiano](https://img.shields.io/badge/lang-IT-blue?style=flat-square) ![Русский](https://img.shields.io/badge/lang-RU-blue?style=flat-square) ![Српски](https://img.shields.io/badge/lang-SR-blue?style=flat-square) ![中文](https://img.shields.io/badge/lang-ZH-blue?style=flat-square)
 
 
   <img src="https://github.com/smkrv/ha-text-ai/blob/main/custom_components/ha_text_ai/icons/logo%402x.png" alt="HA Text AI" style="width: 50%; max-width: 256px; max-height: 128px; aspect-ratio: 2/1; object-fit: contain;"/>
 
-### Advanced AI Integration for [Home Assistant](https://www.home-assistant.io/) with LLM multi-provider support
+### Multi-provider LLM integration for [Home Assistant](https://www.home-assistant.io/)
 </div>
 
 <p align="center">
-Transform your smart home experience with powerful AI assistance powered by multiple AI providers including OpenAI GPT, DeepSeek and Anthropic Claude models. Get intelligent responses, automate complex scenarios, and enhance your home automation with advanced natural language processing.
-
+Ask OpenAI, Anthropic Claude, DeepSeek and Google Gemini models questions from your automations and scripts. The integration keeps per-instance conversation history, returns full-length responses through response variables, supports structured JSON output, and exposes token, latency and error metrics as sensor attributes.
 </p>
 
 ---
 
 > [!IMPORTANT]
-> 🤝 Community Driven: for more details on the integration,   
+> Community driven: for more details on the integration,   
 > check out the discussion on the **[Home Assistant Community forum](https://community.home-assistant.io/t/ha-text-ai-transforming-home-automation-through-multi-llm-integration/799741)**
 >
 >  <a href="https://my.home-assistant.io/redirect/hacs_repository/?owner=smkrv&repository=ha-text-ai&category=Integration"><img src="https://my.home-assistant.io/badges/hacs_repository.svg" width="210" height="auto"></a>
 >  
 > [Screenshots](assets/images/screenshots/screenshot.jpg)
 
-## 🌟 Features
+## Features
 
-- 🧠 **Multi-Provider AI Integration**: Support for OpenAI GPT, DeepSeek and Anthropic Claude models
-- 💬 **Advanced Language Processing**: Context-aware, multi-turn conversations
-- 📝 **Enhanced Memory Management**: Secure file-based history storage
-- ⚡ **Performance Optimization**: Efficient token usage and smart rate limiting
-- 🎯 **Advanced Customization**: Per-request model and parameter selection
-- 🔒 **Enhanced Security**: Secure API key management and usage monitoring
-- 🎨 **Improved User Experience**: Intuitive configuration and rich interfaces
-- 🔄 **Automation Integration**: Event-driven responses and template compatibility
+- **Multi-provider support**: OpenAI, Anthropic Claude, DeepSeek, Google Gemini, plus any OpenAI-compatible endpoint
+- **Conversation context**: the model sees previous messages; depth is configurable per request (1-20)
+- **Response variables**: `ask_question` returns the full response directly to the calling automation, bypassing the 255-character state limit
+- **Structured output**: JSON responses matching a schema you provide
+- **Per-request overrides**: model, temperature, max_tokens, system prompt, thinking mode
+- **Usage metrics**: token counters, latency and success/error statistics as sensor attributes
+- **File-based history**: per-instance JSON storage with automatic rotation at 1 MB
 
-<details>
-<summary>📦 Detailed Feature Breakdown</summary>
-
-### 🧠 **Multi-Provider AI Integration**
-- Support for OpenAI GPT models
-- Anthropic Claude integration
-- DeepSeek integration
-- Custom API endpoints
-- Flexible model selection
-
-### 💬 **Advanced Language Processing**
-- Context-aware responses
-- Multi-turn conversations
-- Custom system instructions
-- Natural conversation flow
-
-### 📝 **Enhanced Memory Management**
-- File-based conversation history storage
-- Automatic history rotation
-- Configurable history size limits
-- Secure storage in Home Assistant
-
-### ⚡ **Performance Optimization**
-- Efficient token usage
-- Smart rate limiting
-- Response caching
-- Request interval control
-
-### 🎯 **Advanced Customization**
-- Per-request model selection
-- Adjustable parameters
-- Custom system prompts
-- Temperature control
-
-### 🔒 **Enhanced Security**
-- Secure API key storage
-- Rate limiting protection
-- Error handling
-- Usage monitoring
-
-### 🎨 **Improved User Experience**
-- Intuitive configuration UI
-- Detailed sensor attributes
-- Rich service interface
-- Model selection UI
-
-### 🔄 **Automation Integration**
-- Event-driven responses
-- Conditional logic support
-- Template compatibility
-- Model-specific automation
-
-</details>
-
-#### 🌐 Translations
+#### Translations
 
 | Code | Language | Status |
 |------|----------|--------|
-| 🇩🇪 de | Deutsch | Full |
-| 🇬🇧 en | English | Primary |
-| 🇪🇸 es | Español | Full |
-| 🇮🇳 hi | हिन्दी | Full |
-| 🇮🇹 it | Italiano | Full |
-| 🇷🇺 ru | Русский | Full |
-| 🇷🇸 sr | Српски | Full |
-| 🇨🇳 zh | 中文 | Full |
+| de | Deutsch | Full |
+| en | English | Primary |
+| es | Español | Full |
+| hi | हिन्दी | Full |
+| it | Italiano | Full |
+| ru | Русский | Full |
+| sr | Српски | Full |
+| zh | 中文 | Full |
 
-## 📋 Prerequisites
+## Prerequisites
 
-- Home Assistant 2024.12.0 or later (recommended for best compatibility)
-- Active API key from:
+- Home Assistant 2024.12.0 or later
+- An API key from one of:
   - OpenAI ([Get key](https://platform.openai.com/account/api-keys))
   - Anthropic ([Get key](https://console.anthropic.com/))
   - DeepSeek ([Get key](https://platform.deepseek.com/api_keys))
   - OpenRouter ([Get key](https://openrouter.ai/keys))
-  - Google Gemini 🆕 ([Get key](https://ai.google.dev/gemini-api/docs/api-key)) thanks to ([@Azzedde](https://github.com/Azzedde))
+  - Google Gemini ([Get key](https://ai.google.dev/gemini-api/docs/api-key)) thanks to ([@Azzedde](https://github.com/Azzedde))
   - Any OpenAI-compatible API provider
-- Python 3.9 or newer
-- Stable internet connection
 
 ## Configuration Options
 
-### 🔧 **Core Configuration Settings**
-- 🌐 **API Provider**: OpenAI/Anthropic/DeepSeek/Gemini
-- 🔑 **API Key**: Provider-specific authentication
-- 🤖 **Model Selection**: Flexible, provider-specific models
-- 🌡️ **Temperature**: Creativity control (0.0-2.0)
-- 📏 **Max Tokens**: Response length limit (passed directly to the LLM API to control the maximum length of the response)
-- ⏱️ **Request Interval**: API call throttling
-- 💾 **History Size**: Number of messages to retain
-- 🌍 **Custom API Endpoint**: Optional advanced configuration
+### Core Configuration Settings
+- **API Provider**: OpenAI / Anthropic / DeepSeek / Gemini
+- **API Key**: provider-specific authentication
+- **Model**: any model your provider offers
+- **Temperature**: sampling temperature (0.0-2.0)
+- **Max Tokens**: response length cap, passed to the LLM API
+- **Request Interval**: minimum delay between API calls (seconds)
+- **History Size**: number of conversations to retain
+- **Custom API Endpoint**: for OpenRouter, proxies and self-hosted servers
+- **Disable Thinking**: turn off model reasoning where the provider supports it
+- **Allow Local Network**: permit endpoints on private addresses (needed for local servers like Ollama)
 
-### 🤖 **Recommended Models**
+### Recommended Models
 
 #### OpenAI Models
-- **GPT-5** - The latest flagship model, best for complex reasoning
-- **GPT-5 mini** - A cost-effective and fast model, suitable for most tasks
+- **GPT-5.6 Sol** - flagship tier for the hardest tasks
+- **GPT-5.6 Terra** - mid-tier for high-volume tasks
+- **GPT-5.6 Luna** - fastest and cheapest, enough for most home automation queries
 
-#### Anthropic Claude Models  
-- **Claude Opus 4.1** - The most capable model for handling complex tasks
-- **Claude Sonnet 4** - Offers a balance between performance and cost
-- **Claude Haiku 4** - The fastest and most economical option in the series
+#### Anthropic Claude Models
+- **Claude Fable 5** - the most capable model for complex tasks
+- **Claude Sonnet 5** - balance between quality and cost
+- **Claude Haiku 4.5** - the fastest and cheapest option in the lineup
 
 #### DeepSeek Models
-- **DeepSeek-V3.1** - A general-purpose model for a wide range of tasks
-- **DeepSeek-R1** - A specialized model focused on reasoning and coding
+- **deepseek-v4-flash** - fast general-purpose model (default)
+- **deepseek-v4-pro** - stronger at reasoning and coding
+
+> The legacy model names `deepseek-chat` and `deepseek-reasoner` stop working on 2026-07-24. If your instance still uses one of them, switch the model in the integration options.
 
 #### Google Gemini Models
-- **Gemini 2.5 Pro & 2.5 Flash** - The newest and most advanced models available
-- **Gemini 2.0 Pro & 2.0 Flash** - Previous generation models that are still powerful and efficient
+- **gemini-3.5-flash** - default; Google's strongest currently available model
+- **gemini-3.1-pro** - previous flagship, still supported
+
+> Google shut down `gemini-2.0-flash` on 2026-06-01 and retires the 2.5 family on 2026-10-16. If your instance uses one of those, switch the model in the integration options.
 
 <details>
-<summary>🌐 Potentially Compatible Providers</summary>
+<summary>Potentially Compatible Providers</summary>
 
-#### Flexible Provider Ecosystem
-The integration is designed to be flexible and may work with other providers offering OpenAI-compatible APIs:
+Other providers with OpenAI-compatible APIs may work through the custom endpoint option:
 - Groq
 - Together AI
 - Perplexity AI
 - Mistral AI
-- Google AI
-- Local AI servers (like Ollama)
+- Local AI servers (like Ollama - enable **Allow Local Network** in the options)
 - Custom OpenAI-compatible endpoints
 
-#### 🚨 Compatibility Notes
-- Not all providers guarantee full compatibility
-- Performance may vary between providers
-- Check individual provider's documentation
-- Ensure your API key has sufficient credits/quota
-
-#### 🔍 Provider Compatibility Requirements
-To be compatible, a provider should support:
-- OpenAI-like REST API structure
-- JSON request/response format
-- Standard authentication method
-- Similar model parameter handling
+Compatibility is not guaranteed. A provider needs an OpenAI-like REST API with JSON request/response format, standard bearer authentication and similar parameter handling. Check the provider's documentation and make sure your API key has sufficient quota.
 
 </details>
 
-## ⚡ Installation
+## Installation
 
 ### HACS Installation (Recommended)
 >[!TIP]
@@ -184,10 +120,9 @@ To be compatible, a provider should support:
 
 <a href="https://my.home-assistant.io/redirect/hacs_repository/?owner=smkrv&repository=ha-text-ai&category=Integration"><img src="https://my.home-assistant.io/badges/hacs_repository.svg" width="170" height="auto"></a>
 1. Open HACS in Home Assistant
-2. Click on "Integrations"
-3. Search for "HA Text AI"
-4. Click "Download"
-5. Restart Home Assistant
+2. Search for "HA Text AI"
+3. Click "Download"
+4. Restart Home Assistant
 
 **Alternative Method (Custom Repository):**
 If the integration is not found in the default repository:
@@ -198,119 +133,102 @@ If the integration is not found in the default repository:
 5. Click "Download"
 
 ### Manual Installation
-1. Download the latest release
-2. Extract and copy `custom_components/ha_text_ai` to your `custom_components` directory
+1. Download `ha_text_ai.zip` from the latest release
+2. Extract the archive and copy the `ha_text_ai` folder into your `custom_components` directory
 3. Restart Home Assistant
-4. Add configuration via UI or YAML
+4. Add configuration via UI (Settings > Devices & Services > Add Integration)
 
-## ⚙️ Configuration
+## Configuration
 
 ### Via UI (Recommended)
-1. Go to Settings → Devices & Services
+1. Go to Settings > Devices & Services
 2. Click "Add Integration"
 3. Search for "HA Text AI"
 4. Follow the configuration steps
 
-<details>
-<summary>📦 Via YAML (Advanced)</summary>
+> **Note:** This integration is configured exclusively through the UI (config entries). YAML configuration is not supported.
 
-### Platform Configuration (Global Settings)
+## Quick Start
+
+After configuration you get one entity per instance, named `sensor.ha_text_ai_<name>`. It is a status sensor: it shows the last response and usage metrics, but you don't type questions into it. Questions go through the `ha_text_ai.ask_question` action, called from Developer Tools, automations, or scripts.
+
+### First question, no YAML
+
+1. Open Developer Tools > Actions (called "Services" in older HA versions).
+2. Search for "HA Text AI: Ask Question".
+3. Pick your instance, type a question, press "Perform action".
+4. The response appears below the form.
+
+### In an automation
+
+1. Go to Settings > Automations & scenes > Create automation.
+2. Add a trigger: a button press, a time, a state change.
+3. Add action > search "HA Text AI: Ask Question" > fill in the question and pick your instance.
+4. To use the reply in a follow-up step, the action needs `response_variable: ai_response`. If the visual editor doesn't show a field for it, open the three-dot menu on that action, choose "Edit in YAML", and add the line at the end.
+5. In the next action, `{{ ai_response.response_text }}` holds the full answer, for example as a notification message.
+
+Complete working automations: [Automation Examples](#automation-examples-with-response-variables).
+
+### On a dashboard
+
+The sensor keeps the last question and answer as attributes, so a Markdown card can show them:
 
 ```yaml
-ha_text_ai:
-  api_provider: openai  # Required
-  api_key: !secret ai_api_key  # Required
-  model: gpt-4o  # Strongly recommended
-  temperature: 0.7  # Optional
-  max_tokens: 1000  # Optional
-  request_interval: 1.0  # Optional
-  api_endpoint: https://api.openai.com/v1  # Required
-  system_prompt: |  # Optional
-    You are a home automation expert assistant.
-    Focus on practical and efficient solutions.
+type: markdown
+content: >-
+  **Q:** {{ state_attr('sensor.ha_text_ai_my_assistant', 'question') }}
+
+  **A:** {{ state_attr('sensor.ha_text_ai_my_assistant', 'response') }}
 ```
 
-### Sensor Configuration
+The attributes fill in after the first question. They are capped at 2048 characters, so long answers come back complete only via `response_variable`.
 
-```yaml
-sensor:
-  - platform: ha_text_ai
-    name: "My AI Assistant"  # Required, unique identifier
-    api_provider: openai  # Optional (inherits from platform)
-    model: "gpt-4o"  # Optional
-    temperature: 0.7  # Optional
-    max_tokens: 1000  # Optional
-```
+## Available Services
 
-### 📋 Configuration Parameters
+### Response Variables
 
-#### Platform Configuration
-
-| Parameter | Type | Required | Default | Description |
-|-----------|------|----------|---------|-------------|
-| `api_provider` | String | ✅ | - | AI service provider (openai, anthropic, deepseek, gemini) |
-| `api_key` | String | ✅ | - | Authentication key for AI service |
-| `model` | String | ⚠️ | gpt-4o-mini | Strongly recommended: Specific AI model to use. Default varies by provider |
-| `temperature` | Float | ❌ | 0.1 | Response creativity level (0.0-2.0) |
-| `max_tokens` | Integer | ❌ | 1000 | Maximum response length |
-| `request_interval` | Float | ❌ | 1.0 | Delay between API requests |
-| `api_endpoint` | URL | ⚠️ | Provider default | Custom API endpoint |
-| `system_prompt` | String | ❌ | - | Default context for AI interactions |
-| `max_history_size` | Integer | ❌ | 50 | Maximum number of conversation entries to store |
-| `context_messages` | Integer | ❌ | 5 | Number of previous messages to include in context (1-20) |
-
-#### Sensor Configuration
-
-| Parameter | Type | Required | Default | Description |
-|-----------|------|----------|---------|-------------|
-| `platform` | String | ✅ | - | Must be `ha_text_ai` |
-| `name` | String | ✅ | - | Unique sensor identifier |
-| `api_provider` | String | ❌ | Platform setting | Override global provider |
-| `model` | String | ⚠️ | Provider default | Recommended: Override global model (gpt-4o-mini, deepseek-chat, gemini-2.0-flash) |
-| `temperature` | Float | ❌ | 0.1 | Override global temperature |
-| `max_tokens` | Integer | ❌ | 1000 | Override global max tokens |
-
-</details>
-
-## 🛠️ Available Services
-
-### 🔄 Response Variables (New!)
-
-**HA Text AI now supports response variables** - a powerful feature that returns AI responses directly from service calls, eliminating the need for separate text sensors and the 255-character limitation!
-
-#### ✨ Key Benefits:
-- **Unlimited response length** - No more 255-character truncation
-- **Direct data access** - Get responses immediately in automations
-- **Race condition prevention** - Eliminates conflicts in parallel automations
-- **Simplified workflows** - No need to read from sensors
+`ask_question` returns its result directly to the calling automation via `response_variable`. The full response text comes back regardless of length (no 255-character truncation), it is available immediately without polling sensor state, and each service call gets its own result, so parallel automations don't overwrite each other.
 
 ### ask_question
 ```yaml
 service: ha_text_ai.ask_question
 data:
   question: "What's the optimal temperature for sleeping?"
-  model: "claude-3.5-sonnet"  # optional
+  instance: sensor.ha_text_ai_claude
+  model: "claude-sonnet-5"  # optional, overrides the configured model
   temperature: 0.5  # optional
   max_tokens: 500  # optional
-  context_messages: 10  #optional, number of previous messages to include in context, default: 5
+  context_messages: 10  # optional, previous messages to include (1-20, default 5)
   system_prompt: "You are a sleep optimization expert"  # optional
-  instance: sensor.ha_text_ai_gpt
-response_variable: ai_response  # NEW! Store response data directly
+  disable_thinking: true  # optional, disable model reasoning for this request
+response_variable: ai_response
 ```
 
-#### 📊 Response Data Structure:
+For structured JSON output, add `structured_output` with a schema:
+```yaml
+service: ha_text_ai.ask_question
+data:
+  question: "Suggest three energy-saving actions for tonight"
+  instance: sensor.ha_text_ai_gpt
+  structured_output: true
+  json_schema: >-
+    {"type": "object", "properties": {"actions": {"type": "array", "items": {"type": "string"}}}}
+response_variable: ai_response
+```
+
+#### Response Data Structure
 ```yaml
 # The service returns structured data:
 response_text: "The optimal sleeping temperature is 65-68°F (18-20°C)..."
 tokens_used: 150
 prompt_tokens: 50
 completion_tokens: 100
-model_used: "claude-3.5-sonnet"
-instance: "sensor.ha_text_ai_gpt"
+model_used: "claude-sonnet-5"
+instance: "sensor.ha_text_ai_claude"
 question: "What's the optimal temperature for sleeping?"
-timestamp: "2025-02-09T16:57:00.000Z"
+timestamp: "2026-07-09T16:57:00.000Z"
 success: true
-# error: "Error message" (only present if success: false)
+# error and error_type are present only when success is false
 ```
 
 ### set_system_prompt
@@ -337,15 +255,16 @@ data:
 ```yaml
 service: ha_text_ai.get_history
 data:
-  limit: 5  # optional, number of conversations to return (1-100)
+  limit: 5  # optional, number of conversations to return (values above 200 are clamped); omit to get the full stored history
   filter_model: "gpt-4o"  # optional, filter by specific AI model
-  start_date: "2025-02-01"  # optional, filter conversations from this date
+  start_date: "2026-02-01"  # optional, filter conversations from this date
   include_metadata: false  # optional, include tokens, response time, etc.
   sort_order: "newest"  # optional, sort order: "newest" or "oldest"
   instance: sensor.ha_text_ai_gpt
+response_variable: history_result  # entries are in history_result.history
 ```
 
-## 🚀 Advanced Automation Examples with Response Variables
+## Automation Examples with Response Variables
 
 ### Example 1: Smart Home Advice with Direct Response
 ```yaml
@@ -362,12 +281,12 @@ automation:
         response_variable: ai_advice
       - service: notify.mobile_app
         data:
-          title: "🏠 Smart Home Tip"
+          title: "Smart Home Tip"
           message: |
             {{ ai_advice.response_text }}
             
-            📊 Tokens used: {{ ai_advice.tokens_used }}
-            🤖 Model: {{ ai_advice.model_used }}
+            Tokens used: {{ ai_advice.tokens_used }}
+            Model: {{ ai_advice.model_used }}
 ```
 
 ### Example 2: Weather-Based AI Recommendations
@@ -393,7 +312,7 @@ automation:
         then:
           - service: persistent_notification.create
             data:
-              title: "❄️ Winter Preparation Advice"
+              title: "Winter Preparation Advice"
               message: |
                 {{ winter_advice.response_text }}
                 
@@ -401,7 +320,7 @@ automation:
         else:
           - service: persistent_notification.create
             data:
-              title: "⚠️ AI Service Error"
+              title: "AI Service Error"
               message: "Failed to get winter advice: {{ winter_advice.error }}"
 ```
 
@@ -437,10 +356,10 @@ automation:
           instance: sensor.ha_text_ai_gpt
         response_variable: recommendations
       
-      # Step 3: Send comprehensive report
+      # Step 3: Send the combined report
       - service: notify.telegram
         data:
-          title: "🏠 Home Analysis Report"
+          title: "Home Analysis Report"
           message: |
             **Analysis:**
             {{ status_analysis.response_text }}
@@ -454,11 +373,11 @@ automation:
             - Generated: {{ recommendations.timestamp }}
 ```
 
-### 💡 Migration from Sensors to Response Variables
+### Migration from Sensors to Response Variables
 
-#### Old Method (Limited):
+#### Old Method:
 ```yaml
-# ❌ Old way - limited to 255 characters, race conditions
+# Old way: delay-based polling, response truncated by the 255-character state limit
 automation:
   - alias: "Old AI Response Method"
     action:
@@ -469,12 +388,12 @@ automation:
       - delay: "00:00:05"  # Wait for sensor update
       - service: notify.mobile
         data:
-          message: "{{ state_attr('sensor.ha_text_ai_gpt', 'response')[:255] }}..."  # Truncated!
+          message: "{{ state_attr('sensor.ha_text_ai_gpt', 'response')[:255] }}..."  # Truncated
 ```
 
-#### New Method (Unlimited):
+#### New Method:
 ```yaml
-# ✅ New way - unlimited length, immediate access, no race conditions
+# New way: full response, available immediately
 automation:
   - alias: "New AI Response Method"
     action:
@@ -482,24 +401,23 @@ automation:
         data:
           question: "Long question here..."
           instance: sensor.ha_text_ai_gpt
-        response_variable: ai_response  # Direct access!
+        response_variable: ai_response
       - service: notify.mobile
         data:
-          message: "{{ ai_response.response_text }}"  # Full response, no truncation!
+          message: "{{ ai_response.response_text }}"  # Full response, no truncation
 ```
 
-### 🏷️ HA Text AI Sensor Naming Convention
+### HA Text AI Sensor Naming Convention
 
-#### Character Restrictions  
-- Only lowercase letters (a-z)  
-- Numbers (0-9)  
-- Underscore (_)  
-- Maximum length: 50 characters (including `ha_text_ai_`)
+#### Naming Rules
+- Only lowercase letters (a-z), numbers (0-9) and underscore (_)
+- The part after the `sensor.ha_text_ai_` prefix is limited to 50 characters
+- No spaces; keep it descriptive but short
 
 #### Sensor Name Structure
 ```yaml
 # Always starts with 'sensor.ha_text_ai_'
-# You define only the part after the underscore
+# You define only the part after the prefix
 sensor.ha_text_ai_YOUR_UNIQUE_SUFFIX
 
 # Examples:
@@ -530,110 +448,99 @@ automation:
             {{ state_attr('sensor.ha_text_ai_gpt', 'response') }}
 ```
 
-### 💡 Naming Rules
-- Prefix is always `sensor.ha_text_ai_`
-- Add your unique identifier after the underscore
-- Use lowercase
-- No spaces allowed
-- Keep it descriptive but concise
+### HA Text AI Sensor Attributes
 
-### 🔍 HA Text AI Sensor Attributes
+- **Model and provider**: current model, API provider, model used for the last response
+- **System status**: processing, rate-limit and endpoint state
+- **Performance metrics**: request success/failure counters and latency statistics
+- **Token usage**: total, prompt and completion token counters as reported by the provider's API
+- **Last interaction**: most recent question, response and timestamp
+- **System health**: error counter, maintenance flag, uptime
 
-- 🤖 **Model and Provider Information**: Tracking current AI model and service provider
-- 🚦 **System Status**: Real-time API and processing readiness
-- 📊 **Performance Metrics**: Request success rates and response times
-- 💬 **Conversation Tracking**: Token usage and interaction history are estimated using a heuristic method based on word count and specific word characteristics, which may differ from actual token usage.
-- 🕒 **Last Interaction Details**: Recent query and response tracking
-- ❤️ **System Health**: Error monitoring and service uptime
+Attributes may be 0 or empty until the first request completes.
 
 <details>
-<summary>📦 Detailed Sensor Attributes</summary>
+<summary>Detailed Sensor Attributes</summary>
 
 #### Model and Provider Information
 ```yaml
-# Name of the AI model currently in use (e.g., latest version of GPT)  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Model') }}           # gpt-4o  
+# Model currently configured for this instance
+{{ state_attr('sensor.ha_text_ai_gpt', 'model') }}         # gpt-4o-mini
 
-# Service provider for the AI model (determines API endpoint and authentication)  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Api provider') }}    # openai  
+# Service provider (determines API endpoint and authentication)
+{{ state_attr('sensor.ha_text_ai_gpt', 'api_provider') }}  # openai
 
-# Previous or alternative model configuration  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Last model') }}      # gpt-4o
+# Model that produced the last response (may differ after a per-request override)
+{{ state_attr('sensor.ha_text_ai_gpt', 'last_model') }}    # gpt-4o-mini
 ```
 
 #### System Status
 ```yaml
-# Current operational readiness of the AI service API  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Api status') }}      # ready  
+# Indicates if a request is currently being processed
+{{ state_attr('sensor.ha_text_ai_gpt', 'is_processing') }}    # false
 
-# Indicates if a request is currently being processed  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Is processing') }}   # false  
+# Shows if the API has hit its request rate limit
+{{ state_attr('sensor.ha_text_ai_gpt', 'is_rate_limited') }}  # false
 
-# Shows if the API has hit its request rate limit  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Is rate limited') }} # false  
-
-# Status of the specific API endpoint being used  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Endpoint status') }} # ready
+# Status of the API endpoint being used
+{{ state_attr('sensor.ha_text_ai_gpt', 'endpoint_status') }}  # ready
 ```
 
 #### Performance Metrics
 ```yaml
-# Total number of successfully completed API requests  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Successful requests') }}   # 0  
+# Number of successfully completed API requests
+{{ state_attr('sensor.ha_text_ai_gpt', 'successful_requests') }}  # 42
 
-# Number of API requests that encountered errors  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Failed requests') }}       # 0  
+# Number of API requests that encountered errors
+{{ state_attr('sensor.ha_text_ai_gpt', 'failed_requests') }}      # 0
 
-# Mean time taken to receive a response from the AI service  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Average latency') }}       # 0  
-
-# Maximum time taken for a single request-response cycle  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Max latency') }}           # 0
+# Average / max / min response time, in seconds
+{{ state_attr('sensor.ha_text_ai_gpt', 'average_latency') }}      # 1.85
+{{ state_attr('sensor.ha_text_ai_gpt', 'max_latency') }}          # 4.2
+{{ state_attr('sensor.ha_text_ai_gpt', 'min_latency') }}          # 0.9
 ```
 
 #### Conversation and Token Usage
 ```yaml
-# Number of previous interactions stored in conversation context  
-{{ state_attr('sensor.ha_text_ai_gpt', 'History size') }}          # 0  
+# Number of entries in the current history file
+{{ state_attr('sensor.ha_text_ai_gpt', 'history_size') }}         # 12
 
-# Total number of tokens used across all interactions  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Total tokens') }}          # 0  
+# Token counters as reported by the provider's API
+{{ state_attr('sensor.ha_text_ai_gpt', 'total_tokens') }}         # 4520
+{{ state_attr('sensor.ha_text_ai_gpt', 'prompt_tokens') }}        # 3100
+{{ state_attr('sensor.ha_text_ai_gpt', 'completion_tokens') }}    # 1420
 
-# Tokens used in the input prompts  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Prompt tokens') }}         # 0  
-
-# Tokens used in the AI's generated responses  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Completion tokens') }}     # 0
-
-# Number of entries in current history file
-{{ state_attr('sensor.ha_text_ai_gpt', 'History size') }}          # 0  
-
-# Last few conversation entries (limited to 1 for performance)
-{{ state_attr('sensor.ha_text_ai_gpt', 'conversation_history') }}  # [...]
+# Last 3 conversation entries, each truncated to 256 characters
+# (full history is available via the get_history service)
+{{ state_attr('sensor.ha_text_ai_gpt', 'conversation_history') }} # [...]
 ```
 
 #### Last Interaction Details
 ```yaml
-# Most recent complete response generated by the AI service  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Response') }}        # Last AI response  
+# Most recent response, truncated to 2048 characters in the attribute
+# (the response_variable path returns the full text)
+{{ state_attr('sensor.ha_text_ai_gpt', 'response') }}        # Last AI response
 
-# The most recently processed user query or prompt  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Question') }}        # Last asked question  
+# The most recently processed question
+{{ state_attr('sensor.ha_text_ai_gpt', 'question') }}        # Last asked question
 
-# Precise moment when the last interaction occurred (useful for tracking and logging)  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Last timestamp') }}  # Timestamp
+# When the last interaction occurred
+{{ state_attr('sensor.ha_text_ai_gpt', 'last_timestamp') }}  # Timestamp
 ```
 
 #### System Health
 ```yaml
-# Cumulative count of all errors encountered during AI service interactions  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Total errors') }}    # 0  
+# Cumulative count of errors across all requests
+{{ state_attr('sensor.ha_text_ai_gpt', 'total_errors') }}    # 0
 
-# Indicates if the AI service is currently undergoing scheduled or emergency maintenance  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Is maintenance') }}  # false  
+# Error message of the last failed request (null after a success)
+{{ state_attr('sensor.ha_text_ai_gpt', 'last_error') }}      # null
 
-# Total continuous operational time of the AI service (in hours or days)  
-{{ state_attr('sensor.ha_text_ai_gpt', 'Uptime') }}          # 547,58
+# Maintenance flag
+{{ state_attr('sensor.ha_text_ai_gpt', 'is_maintenance') }}  # false
+
+# Seconds since the integration instance was set up
+{{ state_attr('sensor.ha_text_ai_gpt', 'uptime') }}          # 547.58
 ```
 
 ### History Storage
@@ -643,41 +550,36 @@ Conversation history stored in `.storage/ha_text_ai_history/` directory:
 - Archived history files are timestamped
 - Default maximum file size: 1MB
 
-### 💡 Pro Tips
-- Always check attribute existence
-- Use these attributes for monitoring and automation
-- Some values might be 0 or empty initially
-
 </details>
 
-## 📘 FAQ
+## FAQ
 
 **Q: Which AI providers are supported?**
-A: OpenAI (GPT models), Anthropic (Claude models), DeepSeek, Google Gemini, and OpenRouter are officially supported, with many other OpenAI-compatible providers working as well.
+A: OpenAI, Anthropic, DeepSeek and Google Gemini are built-in providers. OpenRouter and other OpenAI-compatible services work through the OpenAI provider with a custom endpoint.
 
 **Q: How can I reduce API costs?**
-A: Use gpt-4o-mini or claude-3.5-haiku for most queries, implement caching, and optimize token usage.
+A: Use a cheap fast model (GPT-5.6 Luna, Claude Haiku 4.5, deepseek-v4-flash, gemini-3.5-flash) for routine queries, lower `context_messages`, and cap `max_tokens`.
 
 **Q: Are there limitations on the number of requests?**
-A: Depends on your API provider's plan. We recommend monitoring usage and implementing request throttling via `request_interval` configuration.
+A: Depends on your API provider's plan. Monitor usage via the sensor attributes and throttle calls with the `request_interval` option.
 
 **Q: Can I use custom models?**
 A: Yes, you can configure custom endpoints and use any compatible model by specifying it in the configuration.
 
 **Q: How do I switch between different AI providers?**
-A: Simply change the model parameter in your configuration or service calls to use the desired provider's model.
+A: Each integration instance is bound to one provider. Add a separate instance per provider and pick the instance in your service calls; within an instance you can override the model per request.
 
 **Q: What are the token limits for different models?**
-A: Token limits vary by provider and model. OpenAI's gpt-4o supports up to 128K tokens, Claude 3.5 Sonnet supports up to 200K tokens, while smaller models typically have 8K-32K limits. Check your provider's documentation for specific limits.
+A: Context window sizes vary by provider and model - check your provider's documentation. The `max_tokens` option caps only the response length, not the context window.
 
 **Q: How do I monitor token usage?**
-A: Use the sensor attributes like `Total tokens`, `Prompt tokens`, and `Completion tokens` to track usage. You can also create automations to alert you when usage exceeds certain thresholds.
+A: Use the sensor attributes `total_tokens`, `prompt_tokens` and `completion_tokens`. You can also create automations to alert you when usage exceeds a threshold.
 
 **Q: Is my data secure?**
-A: Yes, your data is secure. The system operates entirely on your local machine, keeping your data under your control. API keys are stored securely and all external communications use encrypted connections.
+A: Conversation history and API keys are stored locally in your Home Assistant instance. Questions and context are sent to the provider you configure over HTTPS; nothing is shared with third parties beyond that provider.
 
 **Q: How do context messages work?**
-A: Context messages allow the AI to remember and reference previous conversation history. By default, 5 previous messages are included, but you can customize this from 1 to 20 messages to control the conversation depth and token usage.
+A: Context messages let the AI reference previous conversation history. By default 5 previous messages are included; you can set 1 to 20 per request to balance conversation depth against token usage.
 
 **Q: Where is conversation history stored?**  
 A: History is stored in files under the `.storage/ha_text_ai_history/` directory, with automatic rotation and size management.
@@ -686,9 +588,9 @@ A: History is stored in files under the `.storage/ha_text_ai_history/` directory
 A: Yes, archived history files are stored with timestamps and can be accessed manually if needed.
 
 **Q: How much history is kept?**  
-A: By default, up to 100 conversations are stored, but this can be configured. Files are automatically rotated when they reach 1MB.
+A: 50 conversations by default, configurable up to 100 in the UI. Files are automatically rotated when they reach 1MB.
 
-## 🤝 Contributing
+## Contributing
 
 Contributions welcome! Please read our [Contributing Guide](CONTRIBUTING.md).
 
@@ -711,12 +613,12 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER   
 DEALINGS IN THE SOFTWARE.  
 
-## 📝 License
+## License
 
 Author: SMKRV
-[CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) - see [LICENSE](LICENSE) for details.
+[MIT License](https://opensource.org/licenses/MIT) - see [LICENSE](LICENSE) for details.
 
-## 💡 Support the Project
+## Support the Project
 
 The best support is:
 - Sharing feedback
@@ -730,16 +632,13 @@ If you want to say thanks financially, you can send a small token of appreciatio
 **USDT Wallet (TRC10/TRC20):**
 `TXC9zYHYPfWUGi4Sv4R1ctTBGScXXQk5HZ`
 
-*Open-source is built by community passion!* 🚀
-
 ---
 
 <div align="center"><img src="https://github.com/smkrv/ha-text-ai/blob/2aaf3405759eb2d97624834594e24ace896131df/assets/images/icons/footer_icon.png" alt="HA Text AI" style="width: 128px; height: auto;"/></div>  
 <div align="center">
 
-Made with ❤️ for the Home Assistant Community
+Made for the Home Assistant Community
 
 [Report Bug](https://github.com/smkrv/ha-text-ai/issues) · [Request Feature](https://github.com/smkrv/ha-text-ai/issues)
 
 </div>
- 

custom_components/ha_text_ai/__init__.py

@@ -1,7 +1,7 @@
 """
 The HA Text AI integration.
 
-@license: CC BY-NC-SA 4.0 International
+@license: MIT (https://opensource.org/licenses/MIT)
 @author: SMKRV
 @github: https://github.com/smkrv/ha-text-ai
 @source: https://github.com/smkrv/ha-text-ai
@@ -9,24 +9,23 @@ The HA Text AI integration.
 from __future__ import annotations
 
 import logging
-import os
-import shutil
-import hashlib
-from datetime import datetime, timedelta
-from typing import Any, Dict, TypeVar
+from typing import Any
+
+import asyncio
 
 import voluptuous as vol
-from async_timeout import timeout
 
 from homeassistant.config_entries import ConfigEntry
-from homeassistant.const import CONF_API_KEY, CONF_NAME, Platform
-from homeassistant.core import HomeAssistant, ServiceCall
+from homeassistant.const import CONF_API_KEY, CONF_NAME, EVENT_HOMEASSISTANT_CLOSE
+from homeassistant.core import HomeAssistant, ServiceCall, SupportsResponse
 from homeassistant.exceptions import ConfigEntryNotReady, HomeAssistantError
 from homeassistant.helpers import config_validation as cv
-from homeassistant.helpers import aiohttp_client
+from homeassistant.util import dt as dt_util
 
 from .coordinator import HATextAICoordinator
 from .api_client import APIClient
+from .utils import create_pinned_session, normalize_name, safe_log_data, validate_endpoint
+from .providers import get_default_endpoint, get_default_model, build_auth_headers
 from .const import (
     DOMAIN,
     PLATFORMS,
@@ -38,19 +37,8 @@ from .const import (
     CONF_API_TIMEOUT,
     CONF_API_PROVIDER,
     CONF_CONTEXT_MESSAGES,
-    API_PROVIDER_OPENAI,
-    API_PROVIDER_ANTHROPIC,
-    API_PROVIDER_DEEPSEEK,
-    API_PROVIDER_GEMINI,
-    DEFAULT_MODEL,
-    DEFAULT_DEEPSEEK_MODEL,
-    DEFAULT_GEMINI_MODEL,
     DEFAULT_TEMPERATURE,
     DEFAULT_MAX_TOKENS,
-    DEFAULT_OPENAI_ENDPOINT,
-    DEFAULT_ANTHROPIC_ENDPOINT,
-    DEFAULT_DEEPSEEK_ENDPOINT,
-    DEFAULT_GEMINI_ENDPOINT,
     DEFAULT_REQUEST_INTERVAL,
     DEFAULT_API_TIMEOUT,
     DEFAULT_CONTEXT_MESSAGES,
@@ -60,22 +48,29 @@ from .const import (
     SERVICE_SET_SYSTEM_PROMPT,
     DEFAULT_MAX_HISTORY,
     CONF_MAX_HISTORY_SIZE,
-    ICONS_SUBDOMAIN,
+    CONF_ALLOW_LOCAL_NETWORK,
+    DEFAULT_ALLOW_LOCAL_NETWORK,
+    CONF_DISABLE_THINKING,
+    DEFAULT_DISABLE_THINKING,
 )
 
 _LOGGER = logging.getLogger(__name__)
 
 CONFIG_SCHEMA = cv.config_entry_only_config_schema(DOMAIN)
-ConfigType = TypeVar("ConfigType", bound=Dict[str, Any])
 
 SERVICE_SCHEMA_ASK_QUESTION = vol.Schema({
     vol.Required("instance"): cv.string,
-    vol.Required("question"): cv.string,
-    vol.Optional("system_prompt"): cv.string,
+    vol.Required("question"): vol.All(cv.string, vol.Length(min=1, max=100000)),
+    vol.Optional("system_prompt"): vol.All(cv.string, vol.Length(max=50000)),
     vol.Optional("model"): cv.string,
-    vol.Optional("temperature"): cv.positive_float,
+    vol.Optional("temperature"): vol.All(
+        vol.Coerce(float), vol.Range(min=0.0, max=2.0)
+    ),
     vol.Optional("max_tokens"): cv.positive_int,
     vol.Optional("context_messages"): cv.positive_int,
+    vol.Optional("structured_output", default=False): cv.boolean,
+    vol.Optional("json_schema"): vol.All(cv.string, vol.Length(max=50000)),
+    vol.Optional("disable_thinking"): cv.boolean,
 })
 
 SERVICE_SCHEMA_SET_SYSTEM_PROMPT = vol.Schema({
@@ -85,7 +80,11 @@ SERVICE_SCHEMA_SET_SYSTEM_PROMPT = vol.Schema({
 
 SERVICE_SCHEMA_GET_HISTORY = vol.Schema({
     vol.Required("instance"): cv.string,
-    vol.Optional("limit"): cv.positive_int,
+    # No default and no schema max: omitting limit returns the full history
+    # (pre-2.5.0 behavior) and oversized values are clamped to
+    # ABSOLUTE_MAX_HISTORY_SIZE in history.async_get_history instead of
+    # failing the whole service call.
+    vol.Optional("limit"): vol.All(cv.positive_int, vol.Range(min=1)),
     vol.Optional("filter_model"): cv.string,
     vol.Optional("start_date"): cv.string,
     vol.Optional("include_metadata"): cv.boolean,
@@ -93,28 +92,43 @@ SERVICE_SCHEMA_GET_HISTORY = vol.Schema({
 })
 
 def get_coordinator_by_instance(hass: HomeAssistant, instance: str) -> HATextAICoordinator:
-    """Get coordinator by instance name."""
+    """Get coordinator by instance name or normalized name.
+
+    Accepts instance_name, normalized_name, or sensor entity_id.
+    """
     if instance.startswith("sensor."):
         instance = instance.replace("sensor.ha_text_ai_", "", 1)
 
+    normalized_input = normalize_name(instance)
+
     for entry_id, coord in hass.data[DOMAIN].items():
-        if isinstance(coord, HATextAICoordinator) and coord.instance_name.lower() == instance.lower():
+        if not isinstance(coord, HATextAICoordinator):
+            continue
+        if (
+            coord.instance_name.lower() == instance.lower()
+            or coord.normalized_name == normalized_input
+        ):
             return coord
 
     raise HomeAssistantError(f"Instance {instance} not found")
 
-def get_file_hash(file_path: str) -> str:
-    """Calculate SHA256 hash of file."""
-    sha256_hash = hashlib.sha256()
-    with open(file_path, "rb") as f:
-        for byte_block in iter(lambda: f.read(4096), b""):
-            sha256_hash.update(byte_block)
-    return sha256_hash.hexdigest()
-
-async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
+async def async_setup(hass: HomeAssistant, config: dict[str, Any]) -> bool:
     """Set up the Home Assistant Text AI component."""
     # Initialize domain data storage
     hass.data.setdefault(DOMAIN, {})
+    _async_register_services(hass)
+    return True
+
+def _async_register_services(hass: HomeAssistant) -> None:
+    """Register domain services; safe to call again after unload.
+
+    Unloading the last config entry unregisters the services, and a config
+    entry reload (every options change does one) runs unload + setup_entry
+    without re-running async_setup — so setup_entry must be able to bring
+    the services back.
+    """
+    if hass.services.has_service(DOMAIN, SERVICE_ASK_QUESTION):
+        return
 
     async def async_ask_question(call: ServiceCall) -> dict:
         """Handle ask_question service with response data."""
@@ -127,6 +141,9 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
                 max_tokens=call.data.get("max_tokens"),
                 system_prompt=call.data.get("system_prompt"),
                 context_messages=call.data.get("context_messages"),
+                structured_output=call.data.get("structured_output", False),
+                json_schema=call.data.get("json_schema"),
+                disable_thinking=call.data.get("disable_thinking"),
             )
             
             # Return structured response data
@@ -152,9 +169,10 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
                 "model_used": call.data.get("model", ""),
                 "instance": call.data["instance"],
                 "question": call.data["question"],
-                "timestamp": datetime.now().isoformat(),
+                "timestamp": dt_util.utcnow().isoformat(),
                 "success": False,
-                "error": str(err)
+                "error": str(err),
+                "error_type": type(err).__name__
             }
 
     async def async_clear_history(call: ServiceCall) -> None:
@@ -164,22 +182,26 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
             await coordinator.async_clear_history()
         except Exception as err:
             _LOGGER.error("Error clearing history: %s", str(err))
-            raise HomeAssistantError(f"Failed to clear history: {str(err)}")
+            raise HomeAssistantError(f"Failed to clear history: {str(err)}") from err
 
-    async def async_get_history(call: ServiceCall) -> list:
+    async def async_get_history(call: ServiceCall) -> dict:
         """Handle get_history service."""
         try:
             coordinator = get_coordinator_by_instance(hass, call.data["instance"])
-            return await coordinator.async_get_history(
+            history = await coordinator.async_get_history(
                 limit=call.data.get("limit"),
                 filter_model=call.data.get("filter_model"),
                 start_date=call.data.get("start_date"),
                 include_metadata=call.data.get("include_metadata", False),
                 sort_order=call.data.get("sort_order", "newest")
             )
+            # HA requires action responses to be dicts. The bare list made
+            # every return_response call fail with a server error, so this
+            # path never worked before and the wrapper breaks no consumer.
+            return {"history": history}
         except Exception as err:
             _LOGGER.error("Error getting history: %s", str(err))
-            raise HomeAssistantError(f"Failed to get history: {str(err)}")
+            raise HomeAssistantError(f"Failed to get history: {str(err)}") from err
 
     async def async_set_system_prompt(call: ServiceCall) -> None:
         """Handle set_system_prompt service."""
@@ -188,7 +210,7 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
             await coordinator.async_set_system_prompt(call.data["prompt"])
         except Exception as err:
             _LOGGER.error("Error setting system prompt: %s", str(err))
-            raise HomeAssistantError(f"Failed to set system prompt: {str(err)}")
+            raise HomeAssistantError(f"Failed to set system prompt: {str(err)}") from err
 
     # Register services
     hass.services.async_register(
@@ -196,7 +218,7 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
         SERVICE_ASK_QUESTION,
         async_ask_question,
         schema=SERVICE_SCHEMA_ASK_QUESTION,
-        supports_response=True
+        supports_response=SupportsResponse.OPTIONAL
     )
 
     hass.services.async_register(
@@ -210,7 +232,8 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
         DOMAIN,
         SERVICE_GET_HISTORY,
         async_get_history,
-        schema=SERVICE_SCHEMA_GET_HISTORY
+        schema=SERVICE_SCHEMA_GET_HISTORY,
+        supports_response=SupportsResponse.OPTIONAL
     )
 
     hass.services.async_register(
@@ -220,75 +243,29 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
         schema=SERVICE_SCHEMA_SET_SYSTEM_PROMPT
     )
 
-    # Handle icons
-    try:
-        source_icon_path = os.path.join(
-            os.path.dirname(__file__),
-            ICONS_SUBDOMAIN,
-            'icon@2x.png'
-        )
-
-        destination_directory = os.path.join(
-            hass.config.path('www'),
-            DOMAIN,
-            ICONS_SUBDOMAIN
-        )
-
-        destination_icon_path = os.path.join(
-            destination_directory,
-            'icon.png'
-        )
-
-        if not os.path.exists(source_icon_path):
-            _LOGGER.error("Source icon not found: %s", source_icon_path)
-            return True
-
-        def create_directory():
-            os.makedirs(destination_directory, exist_ok=True)
-
-        await hass.async_add_executor_job(create_directory)
-
-        should_copy = True
-
-        if os.path.exists(destination_icon_path):
-            source_hash = await hass.async_add_executor_job(get_file_hash, source_icon_path)
-            dest_hash = await hass.async_add_executor_job(get_file_hash, destination_icon_path)
-            should_copy = source_hash != dest_hash
-
-        if should_copy:
-            def copy_file():
-                shutil.copyfile(source_icon_path, destination_icon_path)
-
-            await hass.async_add_executor_job(copy_file)
-            _LOGGER.debug("Icon updated: %s", destination_icon_path)
-
-    except PermissionError as e:
-        _LOGGER.error("Permission denied when managing icons: %s", str(e))
-    except Exception as e:
-        _LOGGER.error("Failed to manage icons: %s", str(e))
-
-    return True
-
 async def async_check_api(session, endpoint: str, headers: dict, provider: str, api_timeout: int = DEFAULT_API_TIMEOUT) -> bool:
-    """Check API availability for different providers."""
+    """Check API availability using provider registry configuration."""
     try:
-        if provider == API_PROVIDER_GEMINI:
-            # Gemini API does not support GET /models for validation, just check key presence
-            if headers.get("Authorization", "").replace("Bearer ", ""):
-                return True
-            else:
-                _LOGGER.error("Gemini API key is missing or empty")
-                return False
-        elif provider == API_PROVIDER_ANTHROPIC:
-            check_url = f"{endpoint}/v1/models"
-        elif provider == API_PROVIDER_DEEPSEEK:
-            check_url = f"{endpoint}/models"
-        else:  # OpenAI
-            check_url = f"{endpoint}/models"
+        from .providers import get_provider_config
+        provider_config = get_provider_config(provider)
+        check_path = provider_config.get("check_path")
 
-        async with timeout(api_timeout):
-            async with session.get(check_url, headers=headers) as response:
-                if response.status in [200, 404]:
+        if check_path is None:
+            # Provider does not support /models check (e.g. Gemini)
+            auth_header = provider_config["auth_header"]
+            auth_value = headers.get(auth_header, "").replace(provider_config.get("auth_prefix", ""), "")
+            if auth_value:
+                return True
+            _LOGGER.error("API key is missing or empty for %s", provider)
+            return False
+
+        check_url = f"{endpoint}{check_path}"
+
+        async with asyncio.timeout(api_timeout):
+            async with session.get(
+                check_url, headers=headers, allow_redirects=False
+            ) as response:
+                if response.status == 200:
                     return True
                 elif response.status == 401:
                     _LOGGER.error("Invalid API key")
@@ -305,24 +282,42 @@ async def async_check_api(session, endpoint: str, headers: dict, provider: str,
 
 async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
     """Set up HA Text AI from a config entry."""
-    _LOGGER.debug(f"Setting up HA Text AI entry: {entry.data}")
+    _LOGGER.debug("Setting up HA Text AI entry: %s", safe_log_data(dict(entry.data)))
 
+    session = None
     try:
-        if CONF_API_PROVIDER not in entry.data:
+        # Get provider from data or options (options takes precedence)
+        config = {**entry.data, **entry.options}
+        api_provider = config.get(CONF_API_PROVIDER)
+
+        if not api_provider:
             _LOGGER.error("API provider not specified")
             raise ConfigEntryNotReady("API provider is required")
 
-        # Get configuration (merge data with options to apply any runtime changes)
-        config = {**entry.data, **entry.options}
-        session = aiohttp_client.async_get_clientsession(hass)
-        api_provider = config.get(CONF_API_PROVIDER)
-        model = config.get(CONF_MODEL, DEFAULT_MODEL)
-        endpoint = config.get(
-            CONF_API_ENDPOINT,
-            DEFAULT_OPENAI_ENDPOINT if api_provider == API_PROVIDER_OPENAI
-            else DEFAULT_ANTHROPIC_ENDPOINT
-        ).rstrip('/')
-        api_key = entry.data[CONF_API_KEY]  # API key stays in data, not in options
+        model = config.get(CONF_MODEL, get_default_model(api_provider))
+        raw_endpoint = config.get(CONF_API_ENDPOINT, get_default_endpoint(api_provider))
+        allow_local = config.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK)
+        if allow_local:
+            _LOGGER.info(
+                "Local network mode enabled for endpoint %s — "
+                "SSRF protection relaxed for self-hosted proxies",
+                raw_endpoint,
+            )
+        try:
+            endpoint, resolved_ips = await validate_endpoint(
+                hass, raw_endpoint, allow_local=allow_local
+            )
+        except ValueError as err:
+            _LOGGER.error("Invalid API endpoint: %s", err)
+            raise ConfigEntryNotReady(f"Invalid API endpoint: {err}") from err
+
+        # Pinned session closes DNS-rebinding TOCTOU and isolates cookies
+        # from other integrations sharing the same endpoint hostname.
+        # The integration owns this session: APIClient.shutdown() closes it
+        # on unload, the except handler below closes it on failed setup.
+        session = create_pinned_session(endpoint, resolved_ips)
+        # API key can now be updated via options
+        api_key = config.get(CONF_API_KEY, entry.data.get(CONF_API_KEY))
         instance_name = entry.data.get(CONF_NAME, entry.entry_id)
         request_interval = config.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL)
         api_timeout = config.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT)
@@ -330,18 +325,9 @@ async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
         temperature = config.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE)
         max_history_size = config.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY)
         context_messages = config.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES)
-        is_anthropic = api_provider == API_PROVIDER_ANTHROPIC
+        disable_thinking = config.get(CONF_DISABLE_THINKING, DEFAULT_DISABLE_THINKING)
 
-        headers = {
-            "Content-Type": "application/json",
-            "Accept": "application/json"
-        }
-
-        if is_anthropic:
-            headers["x-api-key"] = api_key
-            headers["anthropic-version"] = "2023-06-01"
-        else:
-            headers["Authorization"] = f"Bearer {api_key}"
+        headers = build_auth_headers(api_provider, api_key)
 
         if not await async_check_api(session, endpoint, headers, api_provider, api_timeout):
             raise ConfigEntryNotReady("API connection failed")
@@ -355,6 +341,7 @@ async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
             api_provider=api_provider,
             model=model,
             api_timeout=api_timeout,
+            api_key=api_key,
         )
 
         coordinator = HATextAICoordinator(
@@ -363,46 +350,89 @@ async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
             model=model,
             update_interval=request_interval,
             instance_name=instance_name,
+            config_entry=entry,
             max_tokens=max_tokens,
             temperature=temperature,
             max_history_size=max_history_size,
             context_messages=context_messages,
-            is_anthropic=is_anthropic,
             api_timeout=api_timeout,
+            disable_thinking=disable_thinking,
         )
 
-        _LOGGER.debug(f"Created coordinator for {instance_name}")
+        # Initialize coordinator (directories, history, metrics)
+        await coordinator.async_initialize()
+
+        _LOGGER.debug("Created coordinator for %s", instance_name)
 
         # Store coordinator
         hass.data.setdefault(DOMAIN, {})
         hass.data[DOMAIN][entry.entry_id] = coordinator
 
-        _LOGGER.debug(f"Stored coordinator in hass.data[{DOMAIN}][{entry.entry_id}]")
+        # A reload after the last entry was unloaded needs the services back.
+        _async_register_services(hass)
+
+        _LOGGER.debug("Stored coordinator in hass.data[%s][%s]", DOMAIN, entry.entry_id)
 
         # Set up platforms
         await hass.config_entries.async_forward_entry_setups(entry, PLATFORMS)
 
-        _LOGGER.debug(f"Setup completed for {instance_name}")
+        # Register update listener for options changes
+        entry.async_on_unload(entry.add_update_listener(async_update_options))
+
+        # HA Core stop does not unload entries, so close the dedicated
+        # session on the CLOSE event too; unload removes this listener
+        # and closes the session via APIClient.shutdown() instead.
+        async def _async_close_session_on_stop(_event) -> None:
+            if not session.closed:
+                await session.close()
+
+        entry.async_on_unload(
+            hass.bus.async_listen_once(
+                EVENT_HOMEASSISTANT_CLOSE, _async_close_session_on_stop
+            )
+        )
+
+        _LOGGER.debug("Setup completed for %s", instance_name)
 
         return True
 
     except Exception as err:
-        _LOGGER.exception(f"Error setting up HA Text AI: {err}")
+        _LOGGER.exception("Error setting up HA Text AI: %s", err)
+        if session is not None and not session.closed:
+            await session.close()
         raise
 
+async def async_update_options(hass: HomeAssistant, entry: ConfigEntry) -> None:
+    """Handle options update - reload the config entry."""
+    _LOGGER.info("Options updated for %s, reloading integration", entry.title)
+    await hass.config_entries.async_reload(entry.entry_id)
+
 async def async_unload_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
     """Unload a config entry."""
     try:
-        if entry.entry_id in hass.data[DOMAIN]:
-            coordinator = hass.data[DOMAIN][entry.entry_id]
+        unload_ok = await hass.config_entries.async_unload_platforms(entry, PLATFORMS)
+        if unload_ok and entry.entry_id in hass.data[DOMAIN]:
+            coordinator = hass.data[DOMAIN].pop(entry.entry_id)
 
             if hasattr(coordinator.client, 'shutdown'):
                 await coordinator.client.shutdown()
 
             await coordinator.async_shutdown()
-            hass.data[DOMAIN].pop(entry.entry_id)
 
-        return await hass.config_entries.async_unload_platforms(entry, PLATFORMS)
+        # When removing the last config entry, also unregister services and
+        # clear the domain bucket so HA doesn't show stale services in the UI.
+        if not hass.data.get(DOMAIN):
+            hass.data.pop(DOMAIN, None)
+            for service in (
+                SERVICE_ASK_QUESTION,
+                SERVICE_CLEAR_HISTORY,
+                SERVICE_GET_HISTORY,
+                SERVICE_SET_SYSTEM_PROMPT,
+            ):
+                if hass.services.has_service(DOMAIN, service):
+                    hass.services.async_remove(DOMAIN, service)
+
+        return unload_ok
 
     except Exception as ex:
         _LOGGER.exception("Error unloading entry: %s", str(ex))

custom_components/ha_text_ai/api_client.py

@@ -1,19 +1,20 @@
 """
 API Client for HA Text AI.
 
-@license: CC BY-NC-SA 4.0 International
+@license: MIT (https://opensource.org/licenses/MIT)
 @author: SMKRV
 @github: https://github.com/smkrv/ha-text-ai
 @source: https://github.com/smkrv/ha-text-ai
 """
+from __future__ import annotations
+
+import asyncio
+import json
 import logging
-import asyncio
-from typing import Any, Dict, List, Optional
+import re
+from typing import Any
 from aiohttp import ClientSession, ClientTimeout
-from async_timeout import timeout
-from datetime import datetime, timedelta
 
-from homeassistant.core import HomeAssistant
 from homeassistant.exceptions import HomeAssistantError
 from .const import (
     DEFAULT_API_TIMEOUT,
@@ -30,7 +31,6 @@ from .const import (
 
 _LOGGER = logging.getLogger(__name__)
 
-
 class APIClient:
     """API Client for OpenAI and Anthropic."""
 
@@ -38,10 +38,11 @@ class APIClient:
         self,
         session: ClientSession,
         endpoint: str,
-        headers: Dict[str, str],
+        headers: dict[str, str],
         api_provider: str,
         model: str,
         api_timeout: int = DEFAULT_API_TIMEOUT,
+        api_key: str | None = None,
     ) -> None:
         """Initialize API client."""
         self.session = session
@@ -51,6 +52,9 @@ class APIClient:
         self.model = model
         self.api_timeout = api_timeout
         self.timeout = ClientTimeout(total=api_timeout)
+        self._api_key = api_key
+        if self.api_provider == API_PROVIDER_GEMINI and not api_key:
+            raise ValueError("Gemini provider requires api_key parameter")
         self._closed = False
 
     async def __aenter__(self):
@@ -86,94 +90,315 @@ class APIClient:
     async def _make_request(
         self,
         url: str,
-        payload: Dict[str, Any],
-    ) -> Dict[str, Any]:
-        """Make API request with retry logic."""
-        # Log request without sensitive data
+        payload: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Make API request with retry logic for transient errors only.
+
+        Retries on:
+        - asyncio.TimeoutError
+        - HTTP 429 (rate limit) — honors Retry-After header when present
+        - HTTP 502/503/504 (upstream transient errors)
+
+        4xx (other than 429) return immediately — they are not retryable.
+        """
         safe_payload = {k: v for k, v in payload.items() if k not in ['messages', 'system']}
-        _LOGGER.debug(f"API Request: URL={url}, Safe payload: {safe_payload}")
-        
+        _LOGGER.debug("API Request: URL=%s, Safe payload: %s", url, safe_payload)
+
+        retryable_5xx = {502, 503, 504}
+
         for attempt in range(API_RETRY_COUNT):
             try:
-                async with timeout(self.api_timeout):
-                    async with self.session.post(
-                        url,
-                        json=payload,
-                        headers=self.headers,
-                        timeout=self.timeout,
-                    ) as response:
-                        _LOGGER.debug(f"Response status: {response.status}")
-                        if response.status != 200:
-                            error_data = await response.json()
-                            # Log error without sensitive data
-                            safe_error = {k: v for k, v in error_data.items() if k not in ['message', 'details']}
-                            _LOGGER.error(f"API error (status {response.status}): {safe_error}")
-                            raise HomeAssistantError(f"API error: status {response.status}")
+                async with self.session.post(
+                    url,
+                    json=payload,
+                    headers=self.headers,
+                    timeout=self.timeout,
+                    # The session pins DNS to validated IPs; following a
+                    # redirect would resolve a new host past that pin.
+                    allow_redirects=False,
+                ) as response:
+                    _LOGGER.debug("Response status: %s", response.status)
+                    if response.status == 200:
                         return await response.json()
-            except asyncio.TimeoutError:
-                _LOGGER.warning(f"Timeout on attempt {attempt + 1}/{API_RETRY_COUNT}")
+
+                    # Try to get error details
+                    error_data = {}
+                    try:
+                        error_data = await response.json()
+                    except Exception:
+                        error_data = {"raw": await response.text()}
+
+                    # Rate limit — retry with backoff, prefer Retry-After header
+                    if response.status == 429:
+                        _LOGGER.warning(
+                            "Rate limit on attempt %d/%d", attempt + 1, API_RETRY_COUNT
+                        )
+                        if attempt < API_RETRY_COUNT - 1:
+                            retry_after = self._parse_retry_after(
+                                response.headers.get("Retry-After")
+                            )
+                            await asyncio.sleep(retry_after or (2 ** attempt))
+                            continue
+                        raise HomeAssistantError("API rate limit exceeded")
+
+                    # Upstream transient errors — retry with backoff
+                    if response.status in retryable_5xx:
+                        _LOGGER.warning(
+                            "Upstream %d on attempt %d/%d",
+                            response.status, attempt + 1, API_RETRY_COUNT,
+                        )
+                        if attempt < API_RETRY_COUNT - 1:
+                            await asyncio.sleep(2 ** attempt)
+                            continue
+                        raise HomeAssistantError(
+                            f"Upstream error after retries: status {response.status}"
+                        )
+
+                    # Other client/server errors — don't retry
+                    truncated_error = str(error_data)[:512]
+                    _LOGGER.error("API error (status %d): %s", response.status, truncated_error)
+                    raise HomeAssistantError(f"API error: status {response.status}")
+
+            except asyncio.TimeoutError as err:
+                _LOGGER.warning("Timeout on attempt %d/%d", attempt + 1, API_RETRY_COUNT)
                 if attempt == API_RETRY_COUNT - 1:
-                    raise HomeAssistantError("API request timed out")
-                await asyncio.sleep(1 * (attempt + 1))
+                    raise HomeAssistantError("API request timed out") from err
+                await asyncio.sleep(2 ** attempt)
+            except HomeAssistantError:
+                raise
             except Exception as e:
-                _LOGGER.warning(f"API request failed on attempt {attempt + 1}/{API_RETRY_COUNT}: {type(e).__name__}")
+                _LOGGER.warning(
+                    "API request failed on attempt %d/%d: %s",
+                    attempt + 1, API_RETRY_COUNT, type(e).__name__,
+                )
                 if attempt == API_RETRY_COUNT - 1:
                     raise
-                await asyncio.sleep(1 * (attempt + 1))
+                await asyncio.sleep(2 ** attempt)
+
+        raise HomeAssistantError("API request failed after all retries")
+
+    @staticmethod
+    def _parse_retry_after(value: str | None) -> float | None:
+        """Parse Retry-After header (seconds). Caps at 60s to avoid long stalls."""
+        if not value:
+            return None
+        try:
+            seconds = float(value.strip())
+        except (ValueError, AttributeError):
+            return None
+        if seconds <= 0:
+            return None
+        return min(seconds, 60.0)
 
     async def create(
         self,
         model: str,
-        messages: List[Dict[str, str]],
+        messages: list[dict[str, str]],
         temperature: float,
         max_tokens: int,
-    ) -> Dict[str, Any]:
+        structured_output: bool = False,
+        json_schema: str | None = None,
+        disable_thinking: bool = False,
+    ) -> dict[str, Any]:
         """Create completion using appropriate API."""
         try:
             self._validate_parameters(temperature, max_tokens)
 
             if self.api_provider == API_PROVIDER_ANTHROPIC:
                 return await self._create_anthropic_completion(
-                    model, messages, temperature, max_tokens
+                    model, messages, temperature, max_tokens,
+                    structured_output, json_schema, disable_thinking
                 )
             elif self.api_provider == API_PROVIDER_DEEPSEEK:
                 return await self._create_deepseek_completion(
-                    model, messages, temperature, max_tokens
+                    model, messages, temperature, max_tokens,
+                    structured_output, json_schema, disable_thinking
                 )
             elif self.api_provider == API_PROVIDER_GEMINI:
                 return await self._create_gemini_completion(
-                    model, messages, temperature, max_tokens
+                    model, messages, temperature, max_tokens,
+                    structured_output, json_schema, disable_thinking
                 )
             else:
                 return await self._create_openai_completion(
-                    model, messages, temperature, max_tokens
+                    model, messages, temperature, max_tokens,
+                    structured_output, json_schema, disable_thinking
                 )
         except Exception as e:
             _LOGGER.error("API request failed: %s", str(e))
-            raise HomeAssistantError(f"API request failed: {str(e)}")
+            raise HomeAssistantError(f"API request failed: {str(e)}") from e
+
+    # Non-reasoning variants whose names otherwise overlap with the
+    # reasoning prefix set (e.g. "gpt-5-chat-latest" is classic chat).
+    _OPENAI_NON_REASONING_PATTERNS: tuple[str, ...] = ("gpt-5-chat",)
+    _OPENAI_REASONING_REGEX = re.compile(
+        r"^(?:o\d+|gpt-[5-9](?:\.\d+)?)(?:[-_].*)?$"
+    )
+
+    @classmethod
+    def _is_openai_reasoning_model(cls, model: str) -> bool:
+        """Detect OpenAI reasoning models (o-series and GPT-5+ family).
+
+        Reasoning models require max_completion_tokens (not max_tokens),
+        do not accept custom temperature, and use "developer" role instead
+        of "system". Uses a regex so future o5/gpt-6 releases are caught
+        without code change. Explicitly excludes chat-variants
+        (e.g. gpt-5-chat-latest) which are classic chat models.
+        """
+        if not model:
+            return False
+        m = model.strip().lower()
+        # OpenRouter-style "openai/o3" prefix — strip provider namespace.
+        if "/" in m:
+            m = m.rsplit("/", 1)[-1]
+        for non_reasoning in cls._OPENAI_NON_REASONING_PATTERNS:
+            if m.startswith(non_reasoning):
+                return False
+        return bool(cls._OPENAI_REASONING_REGEX.match(m))
+
+    @staticmethod
+    def _convert_system_to_developer(
+        messages: list[dict[str, str]],
+    ) -> list[dict[str, str]]:
+        """Rename role "system" to "developer" for OpenAI reasoning models."""
+        return [
+            {**m, "role": "developer"} if m.get("role") == "system" else m
+            for m in messages
+        ]
+
+    # Matches /no_think only as a standalone soft-switch token (word-bounded),
+    # not when users discuss the concept ("discuss /no_think semantics").
+    _NO_THINK_TOKEN_RE = re.compile(r"(?:^|\s)/no_think(?:\s|$)")
+
+    @classmethod
+    def _apply_no_think_tag(
+        cls,
+        messages: list[dict[str, str]],
+    ) -> list[dict[str, str]]:
+        """Append Qwen-style /no_think soft switch to the last user message.
+
+        Why: Qwen3 reasoning models treat "/no_think" in the last user turn as a
+        request to skip thinking. Non-Qwen models ignore the trailing token
+        harmlessly, so this is safe to apply to all OpenAI-compatible backends.
+        Uses word-boundary regex for dedup so that user content mentioning
+        "/no_think" mid-sentence isn't mistaken for an existing soft switch.
+        """
+        if not messages:
+            return messages
+        patched = [m.copy() for m in messages]
+        for i in range(len(patched) - 1, -1, -1):
+            if patched[i].get("role") == "user":
+                content = patched[i].get("content", "")
+                if not cls._NO_THINK_TOKEN_RE.search(content):
+                    patched[i]["content"] = f"{content.rstrip()} /no_think".lstrip()
+                break
+        return patched
+
+    @staticmethod
+    def _strip_think_blocks(text: str) -> str:
+        """Remove <think>...</think> reasoning blocks from model output.
+
+        Why: Some reasoning models (DeepSeek-R1, Qwen-Thinking) emit chain-of-thought
+        wrapped in <think> tags even when thinking is nominally disabled. Strip them
+        so the final answer stays clean. Handles nested blocks via iterative
+        replacement, and drops dangling opening tags when a response is
+        truncated mid-block.
+        """
+        if not text or "<think>" not in text:
+            return text
+        pattern = re.compile(r"<think>.*?</think>", flags=re.DOTALL)
+        cleaned = text
+        # Iterative pass: each iteration peels one layer of nested tags.
+        # Bounded to 10 iterations to avoid pathological inputs.
+        for _ in range(10):
+            new = pattern.sub("", cleaned)
+            if new == cleaned:
+                break
+            cleaned = new
+        # If a truncated response left a dangling <think> open, drop the rest
+        # from that marker onward to avoid leaking partial reasoning.
+        if "<think>" in cleaned:
+            cleaned = cleaned.split("<think>", 1)[0]
+        return cleaned.strip()
+
+    @staticmethod
+    def _apply_structured_output(
+        payload: dict[str, Any],
+        structured_output: bool,
+        json_schema: str | None,
+    ) -> None:
+        """Apply OpenAI-compatible structured output to payload in-place."""
+        if not (structured_output and json_schema):
+            return
+        try:
+            schema = json.loads(json_schema)
+            payload["response_format"] = {
+                "type": "json_schema",
+                "json_schema": {
+                    "name": "structured_response",
+                    "strict": True,
+                    "schema": schema,
+                },
+            }
+        except json.JSONDecodeError as e:
+            _LOGGER.warning("Invalid JSON schema: %s. Falling back to json_object.", e)
+            payload["response_format"] = {"type": "json_object"}
 
     async def _create_deepseek_completion(
         self,
         model: str,
-        messages: List[Dict[str, str]],
+        messages: list[dict[str, str]],
         temperature: float,
         max_tokens: int,
-    ) -> Dict[str, Any]:
-        """Create completion using DeepSeek API."""
+        structured_output: bool = False,
+        json_schema: str | None = None,
+        disable_thinking: bool = False,
+    ) -> dict[str, Any]:
+        """Create completion using DeepSeek API.
+
+        DeepSeek-reasoner (R1) is a reasoning model: it ignores /no_think
+        (thinking is always on by design) and emits reasoning_content as a
+        separate field alongside content. We skip the no_think append for
+        this model and preserve reasoning_content in the response payload
+        so it's available for logging/debug.
+
+        DeepSeek V4+ (deepseek-v4-flash/-pro) selects thinking mode via a
+        top-level "thinking" request parameter instead of the model name,
+        so /no_think does not apply there.
+        """
         url = f"{self.endpoint}/chat/completions"
+        m_lower = model.lower()
+        is_reasoner = "reasoner" in m_lower
+        is_v4plus = re.search(r"deepseek-v[4-9]", m_lower) is not None
+        final_messages = (
+            self._apply_no_think_tag(messages)
+            if (disable_thinking and not is_reasoner and not is_v4plus)
+            else messages
+        )
         payload = {
             "model": model,
-            "messages": messages,
+            "messages": final_messages,
             "temperature": temperature,
             "max_tokens": max_tokens,
-            "stream": False
+            "stream": False,
         }
+        if disable_thinking and is_v4plus:
+            payload["thinking"] = {"type": "disabled"}
+        self._apply_structured_output(payload, structured_output, json_schema)
 
         data = await self._make_request(url, payload)
+        message = data["choices"][0]["message"]
+        content = message.get("content", "")
+        reasoning = message.get("reasoning_content")
+        if disable_thinking and not is_reasoner:
+            content = self._strip_think_blocks(content)
         return {
             "choices": [
                 {
-                    "message": {"content": data["choices"][0]["message"]["content"]},
+                    "message": {
+                        "content": content,
+                        **({"reasoning_content": reasoning} if reasoning else {}),
+                    },
                 }
             ],
             "usage": {
@@ -186,24 +411,59 @@ class APIClient:
     async def _create_openai_completion(
         self,
         model: str,
-        messages: List[Dict[str, str]],
+        messages: list[dict[str, str]],
         temperature: float,
         max_tokens: int,
-    ) -> Dict[str, Any]:
-        """Create completion using OpenAI API."""
+        structured_output: bool = False,
+        json_schema: str | None = None,
+        disable_thinking: bool = False,
+    ) -> dict[str, Any]:
+        """Create completion using OpenAI API.
+
+        Reasoning models (o-series, gpt-5 family) require a different payload
+        shape: max_completion_tokens instead of max_tokens, no custom
+        temperature, and role "developer" instead of "system". When
+        disable_thinking=True for a reasoning model we set reasoning_effort
+        to "low" to minimize hidden CoT tokens. For classic chat models the
+        Qwen-style /no_think soft switch is appended instead.
+        """
         url = f"{self.endpoint}/chat/completions"
-        payload = {
-            "model": model,
-            "messages": messages,
-            "temperature": temperature,
-            "max_tokens": max_tokens,
-        }
+        is_reasoning = self._is_openai_reasoning_model(model)
+
+        if is_reasoning:
+            prepared_messages = self._convert_system_to_developer(messages)
+            payload: dict[str, Any] = {
+                "model": model,
+                "messages": prepared_messages,
+                "max_completion_tokens": max_tokens,
+            }
+            if disable_thinking:
+                # gpt-5+ supports "minimal" (cheapest, lowest-CoT). o-series
+                # rejects "minimal" and accepts low/medium/high — fall back to "low".
+                effort = "minimal" if model.lower().startswith(("gpt-5", "gpt5")) else "low"
+                payload["reasoning_effort"] = effort
+        else:
+            prepared_messages = (
+                self._apply_no_think_tag(messages) if disable_thinking else messages
+            )
+            payload = {
+                "model": model,
+                "messages": prepared_messages,
+                "temperature": temperature,
+                "max_tokens": max_tokens,
+            }
+        self._apply_structured_output(payload, structured_output, json_schema)
 
         data = await self._make_request(url, payload)
+        content = data["choices"][0]["message"]["content"]
+        # Strip <think> blocks only for classic chat models. Reasoning models
+        # never emit the tags in user-facing content.
+        if disable_thinking and not is_reasoning:
+            content = self._strip_think_blocks(content)
         return {
             "choices": [
                 {
-                    "message": {"content": data["choices"][0]["message"]["content"]},
+                    "message": {"content": content},
                 }
             ],
             "usage": {
@@ -216,10 +476,13 @@ class APIClient:
     async def _create_anthropic_completion(
         self,
         model: str,
-        messages: List[Dict[str, str]],
+        messages: list[dict[str, str]],
         temperature: float,
         max_tokens: int,
-    ) -> Dict[str, Any]:
+        structured_output: bool = False,
+        json_schema: str | None = None,
+        disable_thinking: bool = False,
+    ) -> dict[str, Any]:
         """Create completion using Anthropic API."""
         url = f"{self.endpoint}/v1/messages"
 
@@ -234,21 +497,56 @@ class APIClient:
             else:
                 filtered_messages.append(msg)
 
+        # For Anthropic, add structured output instruction to system prompt.
+        # Validate schema is well-formed JSON before concatenation: untrusted
+        # schema strings (built from templates/webhook data) could otherwise
+        # break out of the JSON fence and rewrite the system instruction.
+        if structured_output and json_schema:
+            try:
+                json.loads(json_schema)
+            except json.JSONDecodeError as err:
+                _LOGGER.warning(
+                    "Anthropic: invalid JSON schema, ignoring structured_output: %s", err
+                )
+            else:
+                schema_instruction = (
+                    f"\n\nIMPORTANT: You MUST respond ONLY with valid JSON that matches "
+                    f"this JSON Schema:\n{json_schema}\n"
+                    f"Do not include any text before or after the JSON. "
+                    f"Do not wrap the JSON in markdown code blocks."
+                )
+                if system_prompt:
+                    system_prompt += schema_instruction
+                else:
+                    system_prompt = schema_instruction.strip()
+                _LOGGER.debug("Anthropic structured output enabled via system prompt")
+
+        # Anthropic accepts temperature in [0, 1], not [0, 2] like OpenAI.
+        # Clip silently to avoid a 400 when a user-set config exceeds the cap.
+        clipped_temp = min(1.0, max(0.0, float(temperature)))
         payload = {
             "model": model,
             "messages": filtered_messages,
             "max_tokens": max_tokens,
-            "temperature": temperature,
+            "temperature": clipped_temp,
         }
 
         if system_prompt:
             payload["system"] = system_prompt
 
         data = await self._make_request(url, payload)
+        # Anthropic returns an array of content blocks; if extended thinking
+        # is ever enabled the first block may be type="thinking". Find the
+        # first text-type block instead of hardcoding index [0].
+        content = ""
+        for block in data.get("content", []):
+            if block.get("type") == "text":
+                content = block.get("text", "")
+                break
         return {
             "choices": [
                 {
-                    "message": {"content": data["content"][0]["text"]},
+                    "message": {"content": content},
                 }
             ],
             "usage": {
@@ -258,22 +556,16 @@ class APIClient:
             },
         }
 
-    async def check_connection(self) -> bool:
-        """Check API connection."""
-        try:
-            await self._make_request(self.endpoint, {"test": "connection"})
-            return True
-        except Exception as e:
-            _LOGGER.error(f"Connection check failed: {str(e)}")
-            return False
-
     async def _create_gemini_completion(
         self,
         model: str,
-        messages: List[Dict[str, str]],
+        messages: list[dict[str, str]],
         temperature: float,
         max_tokens: int,
-    ) -> Dict[str, Any]:
+        structured_output: bool = False,
+        json_schema: str | None = None,
+        disable_thinking: bool = False,
+    ) -> dict[str, Any]:
         """Create completion using Gemini API with google-genai library.
 
         Args:
@@ -281,6 +573,8 @@ class APIClient:
             messages: List of message dictionaries with role and content
             temperature: Sampling temperature between 0.0 and 2.0
             max_tokens: Maximum number of tokens to generate
+            structured_output: Enable JSON structured output mode
+            json_schema: JSON Schema for structured output validation
 
         Returns:
             Dictionary with response content and token usage
@@ -292,8 +586,7 @@ class APIClient:
 
             genai = await asyncio.to_thread(import_genai)
 
-            # Extract API key from headers (Bearer token)
-            api_key = self.headers.get("Authorization", "").replace("Bearer ", "")
+            api_key = self._api_key
 
             def create_client():
                 if self.endpoint and self.endpoint != "https://generativelanguage.googleapis.com/v1beta":
@@ -319,6 +612,15 @@ class APIClient:
                         "parts": [{"text": msg['content']}]
                     })
 
+            # Parse JSON schema if structured output is enabled
+            parsed_schema = None
+            if structured_output and json_schema:
+                try:
+                    parsed_schema = json.loads(json_schema)
+                    _LOGGER.debug("Gemini structured output enabled with schema")
+                except json.JSONDecodeError as e:
+                    _LOGGER.warning("Invalid JSON schema provided: %s. Structured output disabled.", e)
+
             # Create configuration
             def create_config():
                 from google.genai import types
@@ -331,6 +633,35 @@ class APIClient:
                 if system_instruction:
                     config.system_instruction = system_instruction.strip()
 
+                # Add structured output configuration for Gemini
+                if structured_output and parsed_schema:
+                    config.response_mime_type = "application/json"
+                    config.response_schema = parsed_schema
+
+                # Disable thinking. Gemini 3.x+ replaced the numeric
+                # thinking_budget with a semantic thinking_level; Pro
+                # variants do not accept MINIMAL, their floor is LOW.
+                # Gemini 2.5: Flash accepts thinking_budget=0 (fully off),
+                # Pro rejects 0 and requires at least 128 tokens.
+                # 2.0 and earlier ignore the field.
+                if disable_thinking:
+                    m_lower = model.lower()
+                    try:
+                        if re.search(r"gemini-[3-9]", m_lower):
+                            level = "LOW" if "pro" in m_lower else "MINIMAL"
+                            config.thinking_config = types.ThinkingConfig(
+                                thinking_level=level
+                            )
+                        else:
+                            budget = 128 if "2.5-pro" in m_lower else 0
+                            config.thinking_config = types.ThinkingConfig(
+                                thinking_budget=budget
+                            )
+                    except (AttributeError, TypeError, ValueError) as err:
+                        _LOGGER.debug(
+                            "ThinkingConfig not supported by this google-genai version: %s", err
+                        )
+
                 return config
 
             config = await asyncio.to_thread(create_config)
@@ -338,7 +669,6 @@ class APIClient:
             def generate_content():
                 # For single message without history, use generate_content
                 if len(contents) <= 1:
-                    # If we have no content yet, create a simple prompt
                     if not contents:
                         prompt = "I need your assistance."
                     else:
@@ -350,18 +680,35 @@ class APIClient:
                         config=config
                     )
                 else:
-                    # For multi-turn conversations, use chat
-                    chat = client.chats.create(model=model, config=config)
+                    # For multi-turn conversations, pass history to chat
+                    # and only send the last user message
+                    last_user_msg = None
+                    history = []
 
-                    # Send all messages in sequence
-                    for content in contents:
-                        if content["role"] == "user":
-                            response = chat.send_message(content["parts"][0]["text"])
-                            # We don't send assistant messages as they're already part of the history
+                    # Find the last user message — that's the new query
+                    for i in range(len(contents) - 1, -1, -1):
+                        if contents[i]["role"] == "user":
+                            last_user_msg = contents[i]["parts"][0]["text"]
+                            history = contents[:i]
+                            break
 
-                    return response
+                    if last_user_msg is None:
+                        # No user messages at all — shouldn't happen, but handle gracefully
+                        return client.models.generate_content(
+                            model=model,
+                            contents="I need your assistance.",
+                            config=config
+                        )
 
-            response = await asyncio.to_thread(generate_content)
+                    chat = client.chats.create(
+                        model=model, config=config, history=history
+                    )
+                    return chat.send_message(last_user_msg)
+
+            # Gemini uses sync SDK via to_thread, so needs its own timeout
+            # (aiohttp ClientTimeout doesn't apply here)
+            async with asyncio.timeout(self.api_timeout):
+                response = await asyncio.to_thread(generate_content)
 
             # Extract response text
             def extract_response():
@@ -388,6 +735,9 @@ class APIClient:
 
             response_text, usage = await asyncio.to_thread(extract_response)
 
+            if disable_thinking:
+                response_text = self._strip_think_blocks(response_text)
+
             return {
                 "choices": [{
                     "message": {
@@ -398,13 +748,20 @@ class APIClient:
             }
 
         except ImportError as e:
-            _LOGGER.error(f"Google Gemini library not installed: {str(e)}")
-            raise HomeAssistantError(f"Missing dependency: {str(e)}. Please install google-genai.")
+            _LOGGER.error("Google Gemini library not installed: %s", e)
+            raise HomeAssistantError(
+                "Missing dependency: google-genai. Please install it."
+            ) from e
         except Exception as e:
-            _LOGGER.error(f"Gemini API error: {str(e)}")
-            raise HomeAssistantError(f"Gemini API error: {str(e)}")
+            _LOGGER.error("Gemini API error: %s", e)
+            raise HomeAssistantError(f"Gemini API request failed: {e}") from e
 
     async def shutdown(self) -> None:
-        """Shutdown API client."""
+        """Shutdown API client and close its dedicated session."""
         _LOGGER.debug("Shutting down API client")
-        await self.session.close()
+        self._closed = True
+        # The session is dedicated to this config entry (pinned resolver,
+        # isolated cookie jar), so it must be closed here to release the
+        # connector; nothing else owns it.
+        if self.session is not None and not self.session.closed:
+            await self.session.close()

custom_components/ha_text_ai/config_flow.py

@@ -1,21 +1,21 @@
 """
 Config flow for HA text AI integration.
 
-@license: CC BY-NC-SA 4.0 International
+@license: MIT (https://opensource.org/licenses/MIT)
 @author: SMKRV
 @github: https://github.com/smkrv/ha-text-ai
 @source: https://github.com/smkrv/ha-text-ai
 """
+from __future__ import annotations
+
 import logging
-from typing import Any, Dict, Optional
-from datetime import datetime, timedelta
+from typing import Any
 
 import voluptuous as vol
 from homeassistant import config_entries
 from homeassistant.const import CONF_API_KEY, CONF_NAME
 from homeassistant.core import callback
-from homeassistant.data_entry_flow import FlowResult
-from homeassistant.helpers.aiohttp_client import async_get_clientsession
+from homeassistant.config_entries import ConfigFlowResult
 from homeassistant.helpers import selector
 
 from .const import (
@@ -33,17 +33,10 @@ from .const import (
     API_PROVIDER_DEEPSEEK,
     API_PROVIDER_GEMINI,
     API_PROVIDERS,
-    DEFAULT_MODEL,
-    DEFAULT_DEEPSEEK_MODEL,
-    DEFAULT_GEMINI_MODEL,
     DEFAULT_TEMPERATURE,
     DEFAULT_MAX_TOKENS,
     DEFAULT_REQUEST_INTERVAL,
     DEFAULT_API_TIMEOUT,
-    DEFAULT_OPENAI_ENDPOINT,
-    DEFAULT_ANTHROPIC_ENDPOINT,
-    DEFAULT_DEEPSEEK_ENDPOINT,
-    DEFAULT_GEMINI_ENDPOINT,
     DEFAULT_CONTEXT_MESSAGES,
     MIN_TEMPERATURE,
     MAX_TEMPERATURE,
@@ -53,19 +46,62 @@ from .const import (
     MIN_API_TIMEOUT,
     MAX_API_TIMEOUT,
     DEFAULT_NAME_PREFIX,
+    DEFAULT_INSTANCE_NAME,
     DEFAULT_MAX_HISTORY,
     CONF_MAX_HISTORY_SIZE,
+    MIN_CONTEXT_MESSAGES,
+    MAX_CONTEXT_MESSAGES,
+    MIN_HISTORY_SIZE,
+    MAX_HISTORY_SIZE,
+    CONF_ALLOW_LOCAL_NETWORK,
+    DEFAULT_ALLOW_LOCAL_NETWORK,
+    CONF_DISABLE_THINKING,
+    DEFAULT_DISABLE_THINKING,
 )
+from homeassistant.util import dt as dt_util
+
+from .utils import (
+    create_pinned_session,
+    normalize_name,
+    safe_log_data,
+    validate_endpoint,
+)
+from .providers import get_default_endpoint, get_default_model, build_auth_headers
 
 _LOGGER = logging.getLogger(__name__)
 
-
-def normalize_name(name: str) -> str:
-    """Normalize name to conform to HA naming convention using underscores."""
-    normalized = ''.join(c if c.isalnum() or c == '_' else '_' for c in name)
-    normalized = '_'.join(filter(None, normalized.split('_')))
-    return normalized.lower()
-
+def _build_parameter_schema(data: dict[str, Any]) -> dict:
+    """Build shared parameter schema fields used by both ConfigFlow and OptionsFlow."""
+    return {
+        vol.Optional(
+            CONF_TEMPERATURE,
+            default=data.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE),
+        ): vol.All(vol.Coerce(float), vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)),
+        vol.Optional(
+            CONF_MAX_TOKENS,
+            default=data.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS),
+        ): vol.All(vol.Coerce(int), vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)),
+        vol.Optional(
+            CONF_REQUEST_INTERVAL,
+            default=data.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL),
+        ): vol.All(vol.Coerce(float), vol.Range(min=MIN_REQUEST_INTERVAL)),
+        vol.Optional(
+            CONF_API_TIMEOUT,
+            default=data.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT),
+        ): vol.All(vol.Coerce(int), vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)),
+        vol.Optional(
+            CONF_CONTEXT_MESSAGES,
+            default=data.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES),
+        ): vol.All(vol.Coerce(int), vol.Range(min=MIN_CONTEXT_MESSAGES, max=MAX_CONTEXT_MESSAGES)),
+        vol.Optional(
+            CONF_MAX_HISTORY_SIZE,
+            default=data.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY),
+        ): vol.All(vol.Coerce(int), vol.Range(min=MIN_HISTORY_SIZE, max=MAX_HISTORY_SIZE)),
+        vol.Optional(
+            CONF_DISABLE_THINKING,
+            default=data.get(CONF_DISABLE_THINKING, DEFAULT_DISABLE_THINKING),
+        ): bool,
+    }
 
 class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
     """Handle a config flow for HA text AI."""
@@ -78,7 +114,7 @@ class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
         self._data = {}
         self._provider = None
 
-    async def async_step_user(self, user_input: Optional[Dict[str, Any]] = None) -> FlowResult:
+    async def async_step_user(self, user_input: dict[str, Any] | None = None) -> ConfigFlowResult:
         """Handle the initial step."""
         if user_input is None:
             return self.async_show_form(
@@ -96,76 +132,45 @@ class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
         self._provider = user_input[CONF_API_PROVIDER]
         return await self.async_step_provider()
 
-    async def async_step_provider(self, user_input: Optional[Dict[str, Any]] = None) -> FlowResult:
+    def _build_provider_schema(
+        self, data: dict[str, Any] | None = None
+    ) -> vol.Schema:
+        """Build provider configuration schema with optional defaults from data."""
+        defaults = data or {}
+        schema_dict = {
+            vol.Required(CONF_NAME, default=defaults.get(CONF_NAME, DEFAULT_INSTANCE_NAME)): str,
+            vol.Required(CONF_API_KEY): selector.TextSelector(
+                selector.TextSelectorConfig(type=selector.TextSelectorType.PASSWORD)
+            ),
+            vol.Required(CONF_MODEL, default=defaults.get(CONF_MODEL, get_default_model(self._provider))): str,
+            vol.Required(CONF_API_ENDPOINT, default=defaults.get(CONF_API_ENDPOINT, get_default_endpoint(self._provider))): str,
+            vol.Optional(
+                CONF_ALLOW_LOCAL_NETWORK,
+                default=defaults.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK),
+            ): bool,
+        }
+        schema_dict.update(_build_parameter_schema(defaults))
+        return vol.Schema(schema_dict)
+
+    async def async_step_provider(self, user_input: dict[str, Any] | None = None) -> ConfigFlowResult:
         """Handle provider configuration step."""
         self._errors = {}
 
         if user_input is None:
-            # Selecting an endpoint by provider
-            default_endpoint = {
-                API_PROVIDER_OPENAI: DEFAULT_OPENAI_ENDPOINT,
-                API_PROVIDER_ANTHROPIC: DEFAULT_ANTHROPIC_ENDPOINT,
-                API_PROVIDER_DEEPSEEK: DEFAULT_DEEPSEEK_ENDPOINT,
-                API_PROVIDER_GEMINI: DEFAULT_GEMINI_ENDPOINT,
-            }.get(self._provider, DEFAULT_OPENAI_ENDPOINT)
-
-            # Selecting the default model by provider
-            default_model = (
-                DEFAULT_DEEPSEEK_MODEL if self._provider == API_PROVIDER_DEEPSEEK else
-                DEFAULT_GEMINI_MODEL if self._provider == API_PROVIDER_GEMINI else
-                DEFAULT_MODEL
-            )
-
             return self.async_show_form(
                 step_id="provider",
-                data_schema=vol.Schema({
-                    vol.Required(CONF_NAME, default="my_assistant"): str,
-                    vol.Required(CONF_API_KEY): str,
-                    vol.Required(CONF_MODEL, default=default_model): str,
-                    vol.Required(CONF_API_ENDPOINT, default=default_endpoint): str,
-                    vol.Optional(CONF_TEMPERATURE, default=DEFAULT_TEMPERATURE): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-                    ),
-                    vol.Optional(CONF_MAX_TOKENS, default=DEFAULT_MAX_TOKENS): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-                    ),
-                    vol.Optional(CONF_REQUEST_INTERVAL, default=DEFAULT_REQUEST_INTERVAL): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_REQUEST_INTERVAL)
-                    ),
-                    vol.Optional(CONF_API_TIMEOUT, default=DEFAULT_API_TIMEOUT): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-                    ),
-                    vol.Optional(
-                        CONF_CONTEXT_MESSAGES,
-                        default=DEFAULT_CONTEXT_MESSAGES
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=20)
-                    ),
-                    vol.Optional(
-                        CONF_MAX_HISTORY_SIZE,
-                        default=DEFAULT_MAX_HISTORY
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=100)
-                    ),
-                })
+                data_schema=self._build_provider_schema(),
             )
 
-        # Debug log to identify what's in the input
-        _LOGGER.debug(f"Provider step input data: {user_input}")
+        _LOGGER.debug("Provider step input data: %s", safe_log_data(user_input))
 
         input_copy = user_input.copy()
 
         # Check if CONF_NAME exists in input_copy and ensure it's not empty
         if CONF_NAME not in input_copy or not input_copy[CONF_NAME]:
-            _LOGGER.warning(f"Missing name in configuration input: {input_copy}")
-            input_copy[CONF_NAME] = f"gemini_assistant_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
-            _LOGGER.info(f"Auto-generated name: {input_copy[CONF_NAME]}")
+            _LOGGER.warning("Missing name in configuration input: %s", safe_log_data(input_copy))
+            input_copy[CONF_NAME] = f"assistant_{dt_util.utcnow().strftime('%Y%m%d_%H%M%S')}"
+            _LOGGER.info("Auto-generated name: %s", input_copy[CONF_NAME])
 
         # Ensure API key is present
         if CONF_API_KEY not in input_copy or not input_copy[CONF_API_KEY]:
@@ -173,282 +178,127 @@ class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
             _LOGGER.error("API validation error: 'api_key'")
             return self.async_show_form(
                 step_id="provider",
-                data_schema=vol.Schema({
-                    vol.Required(CONF_NAME, default=input_copy.get(CONF_NAME, "my_assistant")): str,
-                    vol.Required(CONF_API_KEY): str,
-                    vol.Required(CONF_MODEL, default=input_copy.get(CONF_MODEL, DEFAULT_GEMINI_MODEL if self._provider == API_PROVIDER_GEMINI else DEFAULT_MODEL)): str,
-                    vol.Required(CONF_API_ENDPOINT, default=input_copy.get(CONF_API_ENDPOINT, DEFAULT_GEMINI_ENDPOINT if self._provider == API_PROVIDER_GEMINI else DEFAULT_OPENAI_ENDPOINT)): str,
-                    vol.Optional(CONF_TEMPERATURE, default=input_copy.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE)): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-                    ),
-                    vol.Optional(CONF_MAX_TOKENS, default=input_copy.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS)): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-                    ),
-                    vol.Optional(CONF_REQUEST_INTERVAL, default=input_copy.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL)): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_REQUEST_INTERVAL)
-                    ),
-                    vol.Optional(CONF_API_TIMEOUT, default=input_copy.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT)): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-                    ),
-                    vol.Optional(
-                        CONF_CONTEXT_MESSAGES,
-                        default=input_copy.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES)
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=20)
-                    ),
-                    vol.Optional(
-                        CONF_MAX_HISTORY_SIZE,
-                        default=input_copy.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY)
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=100)
-                    ),
-                }),
+                data_schema=self._build_provider_schema(input_copy),
                 errors=self._errors
             )
 
         try:
-            # Validate and normalize the name
             normalized_name = self._validate_and_normalize_name(input_copy[CONF_NAME])
             input_copy[CONF_NAME] = normalized_name
         except ValueError as e:
             return self.async_show_form(
                 step_id="provider",
-                data_schema=vol.Schema({
-                    vol.Required(CONF_NAME, default=input_copy.get(CONF_NAME, "my_assistant")): str,
-                    vol.Required(CONF_API_KEY, default=input_copy.get(CONF_API_KEY, "")): str,
-                    vol.Required(CONF_MODEL, default=input_copy.get(CONF_MODEL, DEFAULT_GEMINI_MODEL if self._provider == API_PROVIDER_GEMINI else DEFAULT_MODEL)): str,
-                    vol.Required(CONF_API_ENDPOINT, default=input_copy.get(CONF_API_ENDPOINT, DEFAULT_GEMINI_ENDPOINT if self._provider == API_PROVIDER_GEMINI else DEFAULT_OPENAI_ENDPOINT)): str,
-                    vol.Optional(CONF_TEMPERATURE, default=input_copy.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE)): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-                    ),
-                    vol.Optional(CONF_MAX_TOKENS, default=input_copy.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS)): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-                    ),
-                    vol.Optional(CONF_REQUEST_INTERVAL, default=input_copy.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL)): vol.All(
-                        vol.Coerce(float),
-                        vol.Range(min=MIN_REQUEST_INTERVAL)
-                    ),
-                    vol.Optional(CONF_API_TIMEOUT, default=input_copy.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT)): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-                    ),
-                    vol.Optional(
-                        CONF_CONTEXT_MESSAGES,
-                        default=input_copy.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES)
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=20)
-                    ),
-                    vol.Optional(
-                        CONF_MAX_HISTORY_SIZE,
-                        default=input_copy.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY)
-                    ): vol.All(
-                        vol.Coerce(int),
-                        vol.Range(min=1, max=100)
-                    ),
-                }),
+                data_schema=self._build_provider_schema(input_copy),
                 errors={"name": str(e)}
             )
 
         try:
-            # Special handling for Gemini API validation
-            if self._provider == API_PROVIDER_GEMINI:
-                # For Gemini, we just check if API key is present as there's no simple endpoint to validate
-                if not input_copy.get(CONF_API_KEY):
-                    self._errors["base"] = "invalid_auth"
-                    _LOGGER.error("API validation error: 'api_key'")
-                    return self.async_show_form(
-                        step_id="provider",
-                        data_schema=vol.Schema({
-                            vol.Required(CONF_NAME, default=input_copy.get(CONF_NAME, "my_assistant")): str,
-                            vol.Required(CONF_API_KEY): str,
-                            vol.Required(CONF_MODEL, default=input_copy.get(CONF_MODEL, DEFAULT_GEMINI_MODEL)): str,
-                            vol.Required(CONF_API_ENDPOINT, default=input_copy.get(CONF_API_ENDPOINT, DEFAULT_GEMINI_ENDPOINT)): str,
-                            # Other fields remain the same
-                        }),
-                        errors=self._errors
-                    )
-            else:
-                # For other providers, validate API connection
-                if not await self._async_validate_api(input_copy):
-                    return self.async_show_form(
-                        step_id="provider",
-                        data_schema=vol.Schema({
-                            vol.Required(CONF_NAME, default=input_copy.get(CONF_NAME, "my_assistant")): str,
-                            vol.Required(CONF_API_KEY, default=input_copy.get(CONF_API_KEY, "")): str,
-                            vol.Required(CONF_MODEL, default=input_copy.get(CONF_MODEL, DEFAULT_MODEL)): str,
-                            vol.Required(CONF_API_ENDPOINT, default=input_copy.get(CONF_API_ENDPOINT, DEFAULT_OPENAI_ENDPOINT)): str,
-                            vol.Optional(CONF_TEMPERATURE, default=input_copy.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE)): vol.All(
-                                vol.Coerce(float),
-                                vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-                            ),
-                            vol.Optional(CONF_MAX_TOKENS, default=input_copy.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS)): vol.All(
-                                vol.Coerce(int),
-                                vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-                            ),
-                            vol.Optional(CONF_REQUEST_INTERVAL, default=input_copy.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL)): vol.All(
-                                vol.Coerce(float),
-                                vol.Range(min=MIN_REQUEST_INTERVAL)
-                            ),
-                            vol.Optional(CONF_API_TIMEOUT, default=input_copy.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT)): vol.All(
-                                vol.Coerce(int),
-                                vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-                            ),
-                            vol.Optional(
-                                CONF_CONTEXT_MESSAGES,
-                                default=input_copy.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES)
-                            ): vol.All(
-                                vol.Coerce(int),
-                                vol.Range(min=1, max=20)
-                            ),
-                            vol.Optional(
-                                CONF_MAX_HISTORY_SIZE,
-                                default=input_copy.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY)
-                            ): vol.All(
-                                vol.Coerce(int),
-                                vol.Range(min=1, max=100)
-                            ),
-                        }),
-                        errors=self._errors
-                    )
-        except Exception as e:
-            # Handle any unexpected exceptions during validation
+            if not await self._async_validate_api(input_copy):
+                return self.async_show_form(
+                    step_id="provider",
+                    data_schema=self._build_provider_schema(input_copy),
+                    errors=self._errors
+                )
+        except Exception:
             _LOGGER.exception("Unexpected error during API validation")
             return self.async_show_form(
                 step_id="provider",
-                data_schema=vol.Schema({
-                    vol.Required(CONF_NAME, default=input_copy.get(CONF_NAME, "my_assistant")): str,
-                    vol.Required(CONF_API_KEY, default=input_copy.get(CONF_API_KEY, "")): str,
-                    vol.Required(CONF_MODEL, default=input_copy.get(CONF_MODEL, DEFAULT_GEMINI_MODEL if self._provider == API_PROVIDER_GEMINI else DEFAULT_MODEL)): str,
-                    vol.Required(CONF_API_ENDPOINT, default=input_copy.get(CONF_API_ENDPOINT, DEFAULT_GEMINI_ENDPOINT if self._provider == API_PROVIDER_GEMINI else DEFAULT_OPENAI_ENDPOINT)): str,
-                    # Other fields remain the same
-                }),
-                errors={"base": str(e)}
+                data_schema=self._build_provider_schema(input_copy),
+                errors={"base": "unknown"}
             )
 
-        # All validation passed, create the entry
         return await self._create_entry(input_copy)
 
     def _validate_and_normalize_name(self, name: str) -> str:
-        """
-        Validate and normalize name with detailed error handling.
+        """Validate and normalize name.
+
+        Truncates before uniqueness check to prevent collisions.
 
         Raises:
-            ValueError: If name is invalid
-
-        Returns:
-            Normalized name
+            ValueError: If name is invalid or already exists.
         """
-        if not name:
+        if not name or not name.strip():
             raise ValueError("empty")
 
-        name = name.strip()
-        normalized = ''.join(
-            c if c.isalnum() or c in ' _' else '_'  # Only allow underscores
-            for c in name
-        )
+        normalized = normalize_name(name.strip())[:50]
 
-        normalized = normalized.replace(' ', '_').lower()
+        if not normalized:
+            raise ValueError("empty")
 
         for entry in self._async_current_entries():
             if entry.data.get(CONF_NAME, "") == normalized:
                 raise ValueError("name_exists")
 
-        normalized = normalized[:50]
-
-        if not normalized:
-            raise ValueError("empty")
-
         return normalized
 
-    async def _async_validate_api(self, user_input: Dict[str, Any]) -> bool:
-        """Validate API connection."""
+    async def _async_validate_api(self, user_input: dict[str, Any]) -> bool:
+        """Validate API connection using provider registry."""
         try:
             if CONF_API_KEY not in user_input:
                 _LOGGER.error("API validation error: 'api_key'")
                 self._errors["base"] = "invalid_auth"
                 return False
 
-            session = async_get_clientsession(self.hass)
-            headers = self._get_api_headers(user_input)
-            endpoint = user_input[CONF_API_ENDPOINT].rstrip('/')
+            try:
+                allow_local = user_input.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK)
+                endpoint, resolved_ips = await validate_endpoint(
+                    self.hass, user_input[CONF_API_ENDPOINT], allow_local=allow_local
+                )
+            except ValueError as err:
+                _LOGGER.error("Endpoint validation failed: %s", err)
+                self._errors["base"] = "cannot_connect"
+                return False
 
             if self._provider == API_PROVIDER_GEMINI:
                 if not user_input[CONF_API_KEY]:
                     self._errors["base"] = "invalid_auth"
                     return False
                 return True
-            else:
-                check_url = (
-                    f"{endpoint}/v1/models" if self._provider == API_PROVIDER_ANTHROPIC
-                    else f"{endpoint}/models"
-                )
 
-                async with session.get(check_url, headers=headers) as response:
+            headers = build_auth_headers(self._provider, user_input[CONF_API_KEY])
+
+            from .providers import get_provider_config
+            check_path = get_provider_config(self._provider).get("check_path", "/models")
+            check_url = f"{endpoint}{check_path}"
+
+            # Pinned session ensures the reachability check goes to the same
+            # IP that will later be used by api_client (no DNS rebinding).
+            session = create_pinned_session(endpoint, resolved_ips)
+            try:
+                async with session.get(
+                    check_url, headers=headers, allow_redirects=False
+                ) as response:
                     if response.status == 401:
                         self._errors["base"] = "invalid_auth"
                         return False
-                    elif response.status not in [200, 404]:
+                    elif response.status != 200:
                         self._errors["base"] = "cannot_connect"
                         return False
                     return True
+            finally:
+                await session.close()
 
         except Exception as err:
             _LOGGER.error("API validation error: %s", str(err))
             self._errors["base"] = "cannot_connect"
             return False
 
-    def _get_api_headers(self, user_input: Dict[str, Any]) -> Dict[str, str]:
-        """Get API headers based on provider."""
-        if CONF_API_KEY not in user_input:
-            return {"Content-Type": "application/json"}
-
-        api_key = user_input[CONF_API_KEY]
-
-        if self._provider == API_PROVIDER_ANTHROPIC:
-            return {
-                "x-api-key": api_key,
-                "anthropic-version": "2023-06-01",
-                "Content-Type": "application/json"
-            }
-        elif self._provider == API_PROVIDER_GEMINI:
-            return {
-                "Authorization": f"Bearer {api_key}",
-                "Content-Type": "application/json"
-            }
-        return {
-            "Authorization": f"Bearer {api_key}",
-            "Content-Type": "application/json"
-        }
-
-    async def _create_entry(self, user_input: Dict[str, Any]) -> FlowResult:
-        """Create the config entry with comprehensive data preservation."""
+    async def _create_entry(self, user_input: dict[str, Any]) -> ConfigFlowResult:
+        """Create the config entry with unique_id deduplication."""
         instance_name = user_input[CONF_NAME]
         normalized_name = normalize_name(instance_name)
 
-        unique_id = f"{DOMAIN}_{normalized_name}_{self._provider}".lower()
+        unique_id = f"{DOMAIN}_{normalized_name}_{self._provider}"
+        await self.async_set_unique_id(unique_id)
+        self._abort_if_unique_id_configured()
 
-        default_model = (
-            DEFAULT_DEEPSEEK_MODEL if self._provider == API_PROVIDER_DEEPSEEK else
-            DEFAULT_GEMINI_MODEL if self._provider == API_PROVIDER_GEMINI else
-            DEFAULT_MODEL
-        )
+        default_model = get_default_model(self._provider)
 
         entry_data = {
             CONF_API_PROVIDER: self._provider,
             CONF_NAME: instance_name,
-            "normalized_name": normalized_name,
             CONF_API_KEY: user_input.get(CONF_API_KEY),
             CONF_API_ENDPOINT: user_input.get(CONF_API_ENDPOINT),
-            "unique_id": unique_id,
             CONF_MODEL: user_input.get(CONF_MODEL, default_model),
             CONF_TEMPERATURE: user_input.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE),
             CONF_MAX_TOKENS: user_input.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS),
@@ -456,13 +306,11 @@ class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
             CONF_API_TIMEOUT: user_input.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT),
             CONF_CONTEXT_MESSAGES: user_input.get(CONF_CONTEXT_MESSAGES, DEFAULT_CONTEXT_MESSAGES),
             CONF_MAX_HISTORY_SIZE: user_input.get(CONF_MAX_HISTORY_SIZE, DEFAULT_MAX_HISTORY),
+            CONF_ALLOW_LOCAL_NETWORK: user_input.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK),
+            CONF_DISABLE_THINKING: user_input.get(CONF_DISABLE_THINKING, DEFAULT_DISABLE_THINKING),
         }
 
-        for key, value in user_input.items():
-            if key not in entry_data:
-                entry_data[key] = value
-
-        _LOGGER.debug(f"Creating config entry with data: {entry_data}")
+        _LOGGER.debug("Creating config entry with data: %s", safe_log_data(entry_data))
 
         return self.async_create_entry(
             title=instance_name,
@@ -473,84 +321,201 @@ class HATextAIConfigFlow(config_entries.ConfigFlow, domain=DOMAIN):
     @callback
     def async_get_options_flow(config_entry: config_entries.ConfigEntry) -> config_entries.OptionsFlow:
         """Get the options flow for this handler."""
-        return OptionsFlowHandler(config_entry)
-
+        return OptionsFlowHandler()
 
 class OptionsFlowHandler(config_entries.OptionsFlow):
     """Handle options flow."""
 
-    def __init__(self, config_entry: config_entries.ConfigEntry) -> None:
-        """Initialize options flow."""
-        self.config_entry = config_entry
+    async def _async_validate_api(self, provider: str, api_key: str, endpoint: str, *, allow_local: bool = False) -> bool:
+        """Validate API connection using provider registry."""
+        try:
+            if not api_key:
+                self._errors["base"] = "invalid_auth"
+                return False
 
-    async def async_step_init(self, user_input: Optional[Dict[str, Any]] = None) -> FlowResult:
-        """Manage the options."""
-        if user_input is not None:
-            return self.async_create_entry(title="", data=user_input)
+            try:
+                endpoint, resolved_ips = await validate_endpoint(
+                    self.hass, endpoint, allow_local=allow_local
+                )
+            except ValueError as err:
+                _LOGGER.error("Endpoint validation failed: %s", err)
+                self._errors["base"] = "cannot_connect"
+                return False
 
+            if provider == API_PROVIDER_GEMINI:
+                return True
+
+            headers = build_auth_headers(provider, api_key)
+
+            from .providers import get_provider_config
+            check_path = get_provider_config(provider).get("check_path", "/models")
+            check_url = f"{endpoint}{check_path}"
+
+            session = create_pinned_session(endpoint, resolved_ips)
+            try:
+                async with session.get(
+                    check_url, headers=headers, allow_redirects=False
+                ) as response:
+                    if response.status == 401:
+                        self._errors["base"] = "invalid_auth"
+                        return False
+                    elif response.status != 200:
+                        self._errors["base"] = "cannot_connect"
+                        return False
+                    return True
+            finally:
+                await session.close()
+
+        except Exception as err:
+            _LOGGER.error("API validation error: %s", str(err))
+            self._errors["base"] = "cannot_connect"
+            return False
+
+    async def async_step_init(self, user_input: dict[str, Any] | None = None) -> ConfigFlowResult:
+        """Handle provider selection step."""
+        if not hasattr(self, "_errors"):
+            self._errors: dict[str, str] = {}
+            self._selected_provider: str | None = None
         current_data = {**self.config_entry.data, **self.config_entry.options}
-        provider = current_data.get(CONF_API_PROVIDER)
+        current_provider = current_data.get(CONF_API_PROVIDER, API_PROVIDER_OPENAI)
 
-        default_model = (
-            DEFAULT_DEEPSEEK_MODEL if provider == API_PROVIDER_DEEPSEEK else
-            DEFAULT_GEMINI_MODEL if provider == API_PROVIDER_GEMINI else
-            DEFAULT_MODEL
-        )
+        if user_input is not None:
+            self._selected_provider = user_input.get(CONF_API_PROVIDER, current_provider)
+            return await self.async_step_settings()
 
         return self.async_show_form(
             step_id="init",
             data_schema=vol.Schema({
-                vol.Optional(
-                    CONF_MODEL,
-                    default=current_data.get(CONF_MODEL, default_model)
-                ): str,
-                vol.Optional(
-                    CONF_TEMPERATURE,
-                    default=current_data.get(CONF_TEMPERATURE, DEFAULT_TEMPERATURE)
-                ): vol.All(
-                    vol.Coerce(float),
-                    vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-                ),
-                vol.Optional(
-                    CONF_MAX_TOKENS,
-                    default=current_data.get(CONF_MAX_TOKENS, DEFAULT_MAX_TOKENS)
-                ): vol.All(
-                    vol.Coerce(int),
-                    vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-                ),
-                vol.Optional(
-                    CONF_REQUEST_INTERVAL,
-                    default=current_data.get(CONF_REQUEST_INTERVAL, DEFAULT_REQUEST_INTERVAL)
-                ): vol.All(
-                    vol.Coerce(float),
-                    vol.Range(min=MIN_REQUEST_INTERVAL)
-                ),
-                vol.Optional(
-                    CONF_API_TIMEOUT,
-                    default=current_data.get(CONF_API_TIMEOUT, DEFAULT_API_TIMEOUT)
-                ): vol.All(
-                    vol.Coerce(int),
-                    vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-                ),
-                vol.Optional(
-                    CONF_CONTEXT_MESSAGES,
-                    default=current_data.get(
-                        CONF_CONTEXT_MESSAGES,
-                        DEFAULT_CONTEXT_MESSAGES
+                vol.Required(
+                    CONF_API_PROVIDER,
+                    default=current_provider
+                ): selector.SelectSelector(
+                    selector.SelectSelectorConfig(
+                        options=API_PROVIDERS,
+                        translation_key="api_provider"
                     )
-                ): vol.All(
-                    vol.Coerce(int),
-                    vol.Range(min=1, max=20)
                 ),
-                vol.Optional(
-                    CONF_MAX_HISTORY_SIZE,
-                    default=current_data.get(
-                        CONF_MAX_HISTORY_SIZE,
-                        DEFAULT_MAX_HISTORY
-                    )
-                ): vol.All(
-                    vol.Coerce(int),
-                    vol.Range(min=1, max=100)
-                ),
-            })
+            }),
+            description_placeholders={
+                "current_provider": current_provider
+            }
         )
+
+    async def async_step_settings(self, user_input: dict[str, Any] | None = None) -> ConfigFlowResult:
+        """Handle settings configuration step."""
+        self._errors = {}
+        current_data = {**self.config_entry.data, **self.config_entry.options}
+        provider = self._selected_provider or current_data.get(CONF_API_PROVIDER, API_PROVIDER_OPENAI)
+
+        # Determine if provider changed to show appropriate defaults
+        provider_changed = provider != current_data.get(CONF_API_PROVIDER)
+
+        # Use new defaults if provider changed, otherwise use current values
+        if provider_changed:
+            default_endpoint = get_default_endpoint(provider)
+            default_model = get_default_model(provider)
+        else:
+            default_endpoint = current_data.get(CONF_API_ENDPOINT, get_default_endpoint(provider))
+            default_model = current_data.get(CONF_MODEL, get_default_model(provider))
+
+        if user_input is not None:
+            api_key = user_input.get(CONF_API_KEY, "").strip()
+            endpoint = user_input.get(CONF_API_ENDPOINT, default_endpoint)
+
+            # Require API key re-entry when endpoint or provider changed.
+            # Why: reusing a stored key after provider/endpoint change could
+            # ship credentials to a different service (e.g. OpenAI key to
+            # api.anthropic.com). Always force explicit re-entry.
+            stored_endpoint = current_data.get(CONF_API_ENDPOINT, "")
+            endpoint_changed = endpoint != stored_endpoint
+            if not api_key and (provider_changed or endpoint_changed):
+                self._errors["base"] = "api_key_required"
+                return self.async_show_form(
+                    step_id="settings",
+                    data_schema=self._get_settings_schema(
+                        provider=provider,
+                        current_data=current_data,
+                        user_input=user_input,
+                        default_endpoint=default_endpoint,
+                        default_model=default_model,
+                    ),
+                    errors=self._errors,
+                    description_placeholders={
+                        "provider": provider
+                    }
+                )
+
+            # Fall back to stored key only when neither provider nor endpoint changed.
+            # Defensive: never silently reuse stored key across providers.
+            if not api_key and not provider_changed and not endpoint_changed:
+                api_key = current_data.get(CONF_API_KEY, "")
+
+            allow_local = user_input.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK)
+            if await self._async_validate_api(provider, api_key, endpoint, allow_local=allow_local):
+                final_data = {
+                    CONF_API_PROVIDER: provider,
+                    **user_input,
+                    CONF_API_KEY: api_key,
+                }
+                return self.async_create_entry(title="", data=final_data)
+
+            # Show form again with errors
+            return self.async_show_form(
+                step_id="settings",
+                data_schema=self._get_settings_schema(
+                    provider=provider,
+                    current_data=current_data,
+                    user_input=user_input,
+                    default_endpoint=default_endpoint,
+                    default_model=default_model,
+                ),
+                errors=self._errors,
+                description_placeholders={
+                    "provider": provider
+                }
+            )
+
+        return self.async_show_form(
+            step_id="settings",
+            data_schema=self._get_settings_schema(
+                provider=provider,
+                current_data=current_data,
+                user_input=None,
+                default_endpoint=default_endpoint,
+                default_model=default_model,
+            ),
+            description_placeholders={
+                "provider": provider
+            }
+        )
+
+    def _get_settings_schema(
+        self,
+        provider: str,
+        current_data: dict[str, Any],
+        user_input: dict[str, Any] | None,
+        default_endpoint: str,
+        default_model: str,
+    ) -> vol.Schema:
+        """Build settings schema using shared parameter definitions."""
+        data = user_input or current_data
+
+        schema_dict = {
+            vol.Optional(CONF_API_KEY, default=""): selector.TextSelector(
+                selector.TextSelectorConfig(type=selector.TextSelectorType.PASSWORD)
+            ),
+            vol.Required(
+                CONF_API_ENDPOINT,
+                default=data.get(CONF_API_ENDPOINT, default_endpoint),
+            ): str,
+            vol.Required(
+                CONF_MODEL,
+                default=data.get(CONF_MODEL, default_model),
+            ): str,
+            vol.Optional(
+                CONF_ALLOW_LOCAL_NETWORK,
+                default=data.get(CONF_ALLOW_LOCAL_NETWORK, DEFAULT_ALLOW_LOCAL_NETWORK),
+            ): bool,
+        }
+        schema_dict.update(_build_parameter_schema(data))
+        return vol.Schema(schema_dict)

custom_components/ha_text_ai/const.py

@@ -1,23 +1,19 @@
 """
 Constants for the HA text AI integration.
 
-@license: CC BY-NC-SA 4.0 International
+@license: MIT (https://opensource.org/licenses/MIT)
 @author: SMKRV
 @github: https://github.com/smkrv/ha-text-ai
 @source: https://github.com/smkrv/ha-text-ai
 """
-import os
-import json
+from __future__ import annotations
+
 from typing import Final
-import voluptuous as vol
-from homeassistant.const import Platform, CONF_API_KEY, CONF_NAME
-from homeassistant.helpers import config_validation as cv
-import logging
-_LOGGER = logging.getLogger(__name__)
+from homeassistant.const import Platform
 
 # Domain and platforms
 DOMAIN: Final = "ha_text_ai"
-PLATFORMS: list[str] = ["sensor"]
+PLATFORMS: list[Platform] = [Platform.SENSOR]
 
 # Provider configuration
 CONF_API_PROVIDER: Final = "api_provider"
@@ -33,21 +29,7 @@ API_PROVIDERS: Final = [
     API_PROVIDER_GEMINI
 ]
 
-# Read version from manifest.json
-MANIFEST_PATH = os.path.join(os.path.dirname(__file__), "manifest.json")
-try:
-    with open(MANIFEST_PATH) as manifest_file:
-        manifest = json.load(manifest_file)
-        VERSION = manifest.get("version", "unknown")
-except FileNotFoundError:
-    VERSION = "unknown"
-    _LOGGER.warning("manifest.json not found at %s", MANIFEST_PATH)
-except json.JSONDecodeError as err:
-    VERSION = "unknown"
-    _LOGGER.error("Error decoding JSON from manifest.json: %s", err)
-except Exception as err:
-    VERSION = "unknown"
-    _LOGGER.error("Error reading manifest.json: %s", err)
+VERSION: Final = "2.5.1"
 
 # Default endpoints
 DEFAULT_OPENAI_ENDPOINT: Final = "https://api.openai.com/v1"
@@ -64,27 +46,38 @@ CONF_REQUEST_INTERVAL: Final = "request_interval"
 CONF_API_TIMEOUT: Final = "api_timeout"
 CONF_INSTANCE: Final = "instance"
 CONF_MAX_HISTORY_SIZE: Final = "max_history_size"  # Correct constant name
-CONF_IS_ANTHROPIC: Final = "is_anthropic"
 CONF_CONTEXT_MESSAGES: Final = "context_messages"
+CONF_STRUCTURED_OUTPUT: Final = "structured_output"
+CONF_JSON_SCHEMA: Final = "json_schema"
+CONF_ALLOW_LOCAL_NETWORK: Final = "allow_local_network"
+CONF_DISABLE_THINKING: Final = "disable_thinking"
 
-ABSOLUTE_MAX_HISTORY_SIZE = 500
+ABSOLUTE_MAX_HISTORY_SIZE: Final = 200  # Hard cap; UI allows max MAX_HISTORY_SIZE (100)
 MAX_ATTRIBUTE_SIZE = 4 * 1024
 MAX_HISTORY_FILE_SIZE = 1 * 1024 * 1024
-ICONS_SUBDOMAIN = "icons"
-
 # Default values
 DEFAULT_MODEL: Final = "gpt-4o-mini"
-DEFAULT_DEEPSEEK_MODEL: Final = "deepseek-chat"
-DEFAULT_GEMINI_MODEL: Final = "gemini-2.0-flash"
+DEFAULT_ANTHROPIC_MODEL: Final = "claude-sonnet-4-6"
+# deepseek-chat/deepseek-reasoner are discontinued 2026-07-24; V4 models
+# select thinking mode via a request parameter instead of the model name.
+DEFAULT_DEEPSEEK_MODEL: Final = "deepseek-v4-flash"
+# gemini-2.0-flash was shut down 2026-06-01; 2.5-flash follows 2026-10-16.
+DEFAULT_GEMINI_MODEL: Final = "gemini-3.5-flash"
 DEFAULT_TEMPERATURE: Final = 0.1
 DEFAULT_MAX_TOKENS: Final = 1000
 DEFAULT_REQUEST_INTERVAL: Final = 1.0
-DEFAULT_TIMEOUT: Final = 30
 DEFAULT_API_TIMEOUT: Final = 30
 DEFAULT_MAX_HISTORY: Final = 50
 DEFAULT_NAME: Final = "HA Text AI"
 DEFAULT_NAME_PREFIX = "ha_text_ai"
+DEFAULT_INSTANCE_NAME: Final = "my_assistant"
 DEFAULT_CONTEXT_MESSAGES: Final = 5
+DEFAULT_ALLOW_LOCAL_NETWORK: Final = False
+DEFAULT_DISABLE_THINKING: Final = False
+MIN_CONTEXT_MESSAGES: Final = 1
+MAX_CONTEXT_MESSAGES: Final = 20
+MIN_HISTORY_SIZE: Final = 1
+MAX_HISTORY_SIZE: Final = 100
 
 TRUNCATION_INDICATOR = " ... "
 
@@ -99,7 +92,6 @@ MIN_API_TIMEOUT: Final = 5
 MAX_API_TIMEOUT: Final = 600
 
 # API constants
-API_TIMEOUT: Final = 30  # Legacy constant, use CONF_API_TIMEOUT from config
 API_RETRY_COUNT: Final = 3
 
 # Service names
@@ -181,72 +173,3 @@ STATE_DISCONNECTED: Final = "disconnected"
 EVENT_RESPONSE_RECEIVED: Final = f"{DOMAIN}_response_received"
 EVENT_ERROR_OCCURRED: Final = f"{DOMAIN}_error_occurred"
 EVENT_STATE_CHANGED: Final = f"{DOMAIN}_state_changed"
-
-# Service schema constants
-SERVICE_SCHEMA_ASK_QUESTION = vol.Schema({
-    vol.Required(CONF_INSTANCE): cv.string,
-    vol.Required("question"): cv.string,
-    vol.Optional("system_prompt"): cv.string,
-    vol.Optional("model"): cv.string,
-    vol.Optional("temperature"): vol.All(
-        vol.Coerce(float),
-        vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-    ),
-    vol.Optional("max_tokens"): vol.All(
-        vol.Coerce(int),
-        vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-    ),
-    vol.Optional("context_messages"): vol.All(
-        vol.Coerce(int),
-        vol.Range(min=1, max=20)
-    )
-})
-
-SERVICE_SCHEMA_SET_SYSTEM_PROMPT = vol.Schema({
-    vol.Required(CONF_INSTANCE): cv.string,
-    vol.Required("prompt"): cv.string
-})
-
-SERVICE_SCHEMA_GET_HISTORY = vol.Schema({
-    vol.Required(CONF_INSTANCE): cv.string,
-    vol.Optional("limit", default=10): vol.All(
-        vol.Coerce(int),
-        vol.Range(min=1, max=100),
-    ),
-    vol.Optional("filter_model"): cv.string
-})
-
-# Configuration schema
-CONFIG_SCHEMA = vol.Schema({
-    DOMAIN: vol.Schema({
-        vol.Required(CONF_NAME): cv.string,
-        vol.Required(CONF_API_KEY): cv.string,
-        vol.Required(CONF_API_PROVIDER): vol.In(API_PROVIDERS),
-        vol.Optional(CONF_MODEL, default=DEFAULT_MODEL): cv.string,
-        vol.Optional(CONF_TEMPERATURE, default=DEFAULT_TEMPERATURE): vol.All(
-            vol.Coerce(float),
-            vol.Range(min=MIN_TEMPERATURE, max=MAX_TEMPERATURE)
-        ),
-        vol.Optional(CONF_MAX_TOKENS, default=DEFAULT_MAX_TOKENS): vol.All(
-            vol.Coerce(int),
-            vol.Range(min=MIN_MAX_TOKENS, max=MAX_MAX_TOKENS)
-        ),
-        vol.Optional(CONF_API_ENDPOINT): cv.string,
-        vol.Optional(CONF_REQUEST_INTERVAL, default=DEFAULT_REQUEST_INTERVAL): vol.All(
-            vol.Coerce(float),
-            vol.Range(min=MIN_REQUEST_INTERVAL, max=MAX_REQUEST_INTERVAL)
-        ),
-        vol.Optional(CONF_API_TIMEOUT, default=DEFAULT_API_TIMEOUT): vol.All(
-            vol.Coerce(int),
-            vol.Range(min=MIN_API_TIMEOUT, max=MAX_API_TIMEOUT)
-        ),
-        vol.Optional(CONF_MAX_HISTORY_SIZE, default=DEFAULT_MAX_HISTORY): vol.All(  # Correct usage
-            vol.Coerce(int),
-            vol.Range(min=1, max=100),
-        ),
-        vol.Optional(CONF_CONTEXT_MESSAGES, default=DEFAULT_CONTEXT_MESSAGES): vol.All(
-            vol.Coerce(int),
-            vol.Range(min=1, max=20)
-        )
-    })
-}, extra=vol.ALLOW_EXTRA)

custom_components/ha_text_ai/history.py

@@ -0,0 +1,480 @@
+"""
+History management for HA Text AI integration.
+
+@license: MIT (https://opensource.org/licenses/MIT)
+@author: SMKRV
+@github: https://github.com/smkrv/ha-text-ai
+@source: https://github.com/smkrv/ha-text-ai
+"""
+from __future__ import annotations
+
+import json
+import logging
+import os
+import shutil
+import traceback
+from datetime import datetime
+from typing import Any
+
+import aiofiles
+
+from homeassistant.core import HomeAssistant
+from homeassistant.util import dt as dt_util
+
+from .const import (
+    ABSOLUTE_MAX_HISTORY_SIZE,
+    MAX_ATTRIBUTE_SIZE,
+    MAX_HISTORY_FILE_SIZE,
+    TRUNCATION_INDICATOR,
+)
+
+# Per-entry storage cap (32KB per field) to prevent disk exhaustion
+MAX_STORED_FIELD_SIZE = 32 * 1024
+MAX_ARCHIVE_FILES = 3
+
+_LOGGER = logging.getLogger(__name__)
+
+class AsyncFileHandler:
+    """Async context manager for file operations."""
+
+    def __init__(self, file_path: str, mode: str = "a"):
+        self.file_path = file_path
+        self.mode = mode
+
+    async def __aenter__(self):
+        self.file = await aiofiles.open(self.file_path, self.mode)
+        return self.file
+
+    async def __aexit__(self, exc_type, exc_val, exc_tb):
+        await self.file.close()
+
+def _assert_not_symlink(path: str) -> None:
+    """Refuse to operate on a path that resolves to a symlink.
+
+    Why: another component or an attacker with filesystem access could
+    replace our history file with a symlink pointing at arbitrary disk
+    locations. Then os.remove or shutil.move would hit the target
+    instead of our managed file. Check before destructive ops.
+    """
+    if os.path.islink(path):
+        raise OSError(f"Refusing to operate on symlink: {path}")
+
+class HistoryManager:
+    """Manages conversation history for an instance."""
+
+    def __init__(
+        self,
+        hass: HomeAssistant,
+        instance_name: str,
+        normalized_name: str,
+        history_dir: str,
+        max_history_size: int,
+    ) -> None:
+        self.hass = hass
+        self.instance_name = instance_name
+        self.normalized_name = normalized_name
+        self._history_dir = history_dir
+        self.max_history_size = min(
+            max(1, max_history_size), ABSOLUTE_MAX_HISTORY_SIZE
+        )
+        self._history_file = os.path.join(
+            history_dir, f"{normalized_name}_history.json"
+        )
+        self._max_history_file_size = MAX_HISTORY_FILE_SIZE
+        self._conversation_history: list[dict[str, Any]] = []
+
+    @property
+    def conversation_history(self) -> list[dict[str, Any]]:
+        return self._conversation_history
+
+    @property
+    def history_size(self) -> int:
+        return len(self._conversation_history)
+
+    async def async_initialize(self) -> None:
+        """Initialize history: directories, file, migration."""
+        await self._create_history_dir()
+        await self._check_history_directory()
+        await self._initialize_history_file()
+        await self._migrate_history_from_txt_to_json()
+
+    async def _file_exists(self, path: str) -> bool:
+        try:
+            return await self.hass.async_add_executor_job(os.path.exists, path)
+        except Exception as e:
+            _LOGGER.error("Error checking file existence for %s: %s", path, e)
+            return False
+
+    async def _create_history_dir(self) -> None:
+        try:
+            await self.hass.async_add_executor_job(
+                os.makedirs, self._history_dir, 0o755, True
+            )
+        except PermissionError:
+            _LOGGER.error("Permission denied creating history directory: %s", self._history_dir)
+            raise
+        except OSError as e:
+            _LOGGER.error("Error creating history directory %s: %s", self._history_dir, e)
+            raise
+
+    async def _check_history_directory(self) -> None:
+        """Check history directory permissions and writability."""
+        try:
+            test_file_path = os.path.join(self._history_dir, ".write_test")
+            await self.hass.async_add_executor_job(
+                self._sync_test_directory_write, test_file_path
+            )
+        except PermissionError:
+            _LOGGER.error("No write permissions for history directory: %s", self._history_dir)
+        except Exception as e:
+            _LOGGER.error("Error checking history directory: %s", e)
+
+    @staticmethod
+    def _sync_test_directory_write(test_file_path: str) -> None:
+        try:
+            os.makedirs(os.path.dirname(test_file_path), mode=0o755, exist_ok=True)
+            with open(test_file_path, "w") as f:
+                f.write("Permission test")
+            os.remove(test_file_path)
+        except Exception as e:
+            _LOGGER.error("Directory write test failed: %s", e)
+
+    async def _initialize_history_file(self) -> None:
+        """Initialize history file and load existing history."""
+        try:
+            if await self._file_exists(self._history_file):
+                async with AsyncFileHandler(self._history_file, "r") as f:
+                    content = await f.read()
+                    if content:
+                        history = json.loads(content)
+                        if isinstance(history, list):
+                            self._conversation_history = history[
+                                -self.max_history_size :
+                            ]
+                            _LOGGER.debug(
+                                "Loaded %d history entries for %s",
+                                len(self._conversation_history),
+                                self.instance_name,
+                            )
+            else:
+                async with AsyncFileHandler(self._history_file, "w") as f:
+                    await f.write(json.dumps([]))
+
+            await self._check_history_size()
+        except Exception as e:
+            _LOGGER.error("Could not initialize history file: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def update_history(self, question: str, response: dict) -> None:
+        """Update conversation history.
+
+        In-memory history stores full text for context retrieval.
+        On-disk storage caps per-field size to prevent disk exhaustion.
+        Display truncation is handled by get_limited_history().
+        """
+        try:
+            content = response.get("content", "")
+            history_entry = {
+                "timestamp": dt_util.utcnow().isoformat(),
+                "question": question[:MAX_STORED_FIELD_SIZE],
+                "response": content[:MAX_STORED_FIELD_SIZE],
+            }
+
+            self._conversation_history.append(history_entry)
+
+            while len(self._conversation_history) > self.max_history_size:
+                self._conversation_history.pop(0)
+
+            await self._save_history_to_file()
+        except Exception as e:
+            _LOGGER.error("Error updating history: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def _save_history_to_file(self) -> None:
+        """Serialize in-memory history to file with rotation if needed."""
+        try:
+            data = json.dumps(self._conversation_history, indent=2)
+            data_size = len(data.encode("utf-8"))
+
+            if data_size > MAX_HISTORY_FILE_SIZE:
+                await self._rotate_history()
+
+            async with AsyncFileHandler(self._history_file, "w") as f:
+                await f.write(data)
+        except Exception as e:
+            _LOGGER.error("Error writing history file: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def _check_history_size(self) -> None:
+        if len(self._conversation_history) > self.max_history_size:
+            _LOGGER.warning(
+                "History size (%d) exceeds maximum (%d). Trimming...",
+                len(self._conversation_history), self.max_history_size,
+            )
+            self._conversation_history = self._conversation_history[
+                -self.max_history_size :
+            ]
+
+    async def _check_file_size(self, file_path: str) -> int:
+        try:
+            if await self._file_exists(file_path):
+                return await self.hass.async_add_executor_job(
+                    os.path.getsize, file_path
+                )
+            return 0
+        except Exception as e:
+            _LOGGER.error("Error checking file size for %s: %s", file_path, e)
+            return 0
+
+    async def _rotate_history(self) -> None:
+        try:
+            _LOGGER.debug("Starting history rotation for %s", self._history_file)
+            await self._rotate_history_files()
+        except Exception as e:
+            _LOGGER.error("Error rotating history: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def _rotate_history_files(self) -> None:
+        """Rotate history files with size validation."""
+        try:
+            if await self._file_exists(self._history_file):
+                current_size = await self._check_file_size(self._history_file)
+
+                if current_size > MAX_HISTORY_FILE_SIZE:
+                    _LOGGER.info(
+                        "Rotating history file. Current size: %d, Max: %d",
+                        current_size, MAX_HISTORY_FILE_SIZE,
+                    )
+
+                    archive_file = os.path.join(
+                        self._history_dir,
+                        f"{self.normalized_name}_history_{dt_util.utcnow().strftime('%Y%m%d_%H%M%S')}.json",
+                    )
+
+                    await self.hass.async_add_executor_job(
+                        _assert_not_symlink, self._history_file
+                    )
+                    await self.hass.async_add_executor_job(
+                        shutil.move, self._history_file, archive_file
+                    )
+
+                    async with AsyncFileHandler(self._history_file, "w") as f:
+                        await f.write(
+                            json.dumps(
+                                self._conversation_history[
+                                    -self.max_history_size :
+                                ],
+                                indent=2,
+                            )
+                        )
+
+                    _LOGGER.info("History file rotated to: %s", archive_file)
+
+                    # Clean up old archive files, keep only MAX_ARCHIVE_FILES
+                    await self._cleanup_archives()
+        except Exception as e:
+            _LOGGER.error("History rotation failed: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def _cleanup_archives(self) -> None:
+        """Remove old archive files beyond MAX_ARCHIVE_FILES."""
+        try:
+            prefix = f"{self.normalized_name}_history_"
+
+            def find_archives():
+                archives = []
+                for f in os.listdir(self._history_dir):
+                    if f.startswith(prefix) and f.endswith(".json") and f != os.path.basename(self._history_file):
+                        archives.append(os.path.join(self._history_dir, f))
+                archives.sort()
+                return archives
+
+            archives = await self.hass.async_add_executor_job(find_archives)
+            if len(archives) > MAX_ARCHIVE_FILES:
+                for old_file in archives[:-MAX_ARCHIVE_FILES]:
+                    await self.hass.async_add_executor_job(
+                        _assert_not_symlink, old_file
+                    )
+                    await self.hass.async_add_executor_job(os.remove, old_file)
+                    _LOGGER.debug("Removed old archive: %s", old_file)
+        except Exception as e:
+            _LOGGER.warning("Archive cleanup error: %s", e)
+
+    async def _migrate_history_from_txt_to_json(self) -> None:
+        """Migrate old .txt history to .json format."""
+        try:
+            old_history_file = os.path.join(
+                self._history_dir, f"{self.normalized_name}_history.txt"
+            )
+
+            if not await self._file_exists(old_history_file):
+                return
+
+            # Skip migration if JSON history already has entries
+            if self._conversation_history:
+                _LOGGER.debug(
+                    "JSON history already has %d entries for %s, skipping txt migration",
+                    len(self._conversation_history), self.instance_name,
+                )
+                return
+
+            _LOGGER.info(
+                "Found old history file for %s, migrating to JSON", self.instance_name
+            )
+
+            history_entries = []
+            async with AsyncFileHandler(old_history_file, "r") as f:
+                content = await f.read()
+
+            for line in content.split("\n"):
+                if not line or line.startswith("History initialized at:"):
+                    continue
+                try:
+                    parts = line.split(": ", 1)
+                    if len(parts) != 2:
+                        continue
+                    timestamp = parts[0]
+                    content_parts = parts[1].split(" - ")
+                    if len(content_parts) != 2:
+                        continue
+                    question = content_parts[0].replace("Question: ", "")
+                    response = content_parts[1].replace("Response: ", "")
+                    history_entries.append(
+                        {
+                            "timestamp": timestamp,
+                            "question": question,
+                            "response": response,
+                        }
+                    )
+                except Exception as e:
+                    _LOGGER.warning("Error parsing history line: %s. Error: %s", line, e)
+                    continue
+
+            if history_entries:
+                async with AsyncFileHandler(self._history_file, "w") as f:
+                    await f.write(json.dumps(history_entries, indent=2))
+
+                backup_file = old_history_file + ".backup"
+                await self.hass.async_add_executor_job(
+                    shutil.move, old_history_file, backup_file
+                )
+
+                _LOGGER.info(
+                    "Migrated %d entries from txt to JSON for %s. Old file: %s",
+                    len(history_entries), self.instance_name, backup_file,
+                )
+
+                self._conversation_history = history_entries
+        except Exception as e:
+            _LOGGER.error("Error during history migration for %s: %s", self.instance_name, e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def async_clear_history(self) -> None:
+        """Clear conversation history."""
+        try:
+            self._conversation_history = []
+            if await self._file_exists(self._history_file):
+                await self.hass.async_add_executor_job(
+                    _assert_not_symlink, self._history_file
+                )
+                await self.hass.async_add_executor_job(os.remove, self._history_file)
+            _LOGGER.info("History for %s cleared", self.instance_name)
+        except Exception as e:
+            _LOGGER.error("Error clearing history: %s", e)
+            _LOGGER.debug(traceback.format_exc())
+
+    async def async_get_history(
+        self,
+        limit: int | None = None,
+        filter_model: str | None = None,
+        start_date: str | None = None,
+        include_metadata: bool = False,
+        sort_order: str = "newest",
+        default_model: str = "",
+    ) -> list[dict[str, Any]]:
+        """Get conversation history with optional filtering and sorting."""
+        try:
+            history = self._conversation_history.copy()
+
+            if filter_model:
+                history = [
+                    entry for entry in history if entry.get("model") == filter_model
+                ]
+
+            if start_date:
+                try:
+                    start_dt = datetime.fromisoformat(
+                        start_date.replace("Z", "+00:00")
+                    )
+                    history = [
+                        entry
+                        for entry in history
+                        if datetime.fromisoformat(
+                            entry["timestamp"].replace("Z", "+00:00")
+                        )
+                        >= start_dt
+                    ]
+                except (ValueError, KeyError) as e:
+                    _LOGGER.warning("Invalid start_date format: %s. Error: %s", start_date, e)
+
+            if sort_order == "oldest":
+                history.sort(key=lambda x: x.get("timestamp", ""))
+            else:
+                history.sort(key=lambda x: x.get("timestamp", ""), reverse=True)
+
+            # Clamp limit to ABSOLUTE_MAX_HISTORY_SIZE to prevent pathological
+            # caller requests from producing multi-MB service payloads.
+            if limit and limit > 0:
+                effective_limit = min(int(limit), ABSOLUTE_MAX_HISTORY_SIZE)
+                history = history[:effective_limit]
+
+            if include_metadata:
+                enriched = []
+                for entry in history:
+                    enriched_entry = dict(entry)
+                    enriched_entry["metadata"] = {
+                        "entry_size": len(str(entry)),
+                        "question_length": len(entry.get("question", "")),
+                        "response_length": len(entry.get("response", "")),
+                        "model_used": entry.get("model", default_model),
+                        "instance": self.instance_name,
+                    }
+                    enriched.append(enriched_entry)
+                return enriched
+
+            return history
+        except Exception as e:
+            _LOGGER.error("Error getting history: %s", e)
+            return []
+
+    def get_limited_history(self, max_display: int = 5) -> dict[str, Any]:
+        """Get limited conversation history for sensor attributes.
+
+        Returns last `max_display` entries with truncated text for HA state.
+        """
+        recent = self._conversation_history[-max_display:]
+        limited_history = [
+            {
+                "timestamp": entry["timestamp"],
+                "question": self._truncate_text(entry["question"], 4096),
+                "response": self._truncate_text(entry["response"], 4096),
+            }
+            for entry in recent
+        ]
+
+        return {
+            "entries": limited_history,
+            "info": {
+                "total_entries": len(self._conversation_history),
+                "displayed_entries": len(limited_history),
+            },
+        }
+
+    @staticmethod
+    def _truncate_text(text: str, max_length: int = MAX_ATTRIBUTE_SIZE) -> str:
+        """Safely truncate text to maximum length with indicator."""
+        if not text:
+            return ""
+        if len(text) <= max_length:
+            return text
+        return text[:max_length] + TRUNCATION_INDICATOR

custom_components/ha_text_ai/manifest.json

@@ -2,7 +2,6 @@
   "domain": "ha_text_ai",
   "name": "HA Text AI",
   "after_dependencies": ["http"],
-  "bluetooth": [],
   "codeowners": ["@smkrv"],
   "config_flow": true,
   "dependencies": [],
@@ -11,19 +10,10 @@
   "iot_class": "cloud_polling",
   "issue_tracker": "https://github.com/smkrv/ha-text-ai/issues",
   "loggers": ["custom_components.ha_text_ai"],
-  "mqtt": [],
-  "quality_scale": "silver",
   "requirements": [
-    "openai>=1.12.0",
-    "anthropic>=0.8.0",
-    "google-genai>=1.16.0",
-    "aiohttp>=3.8.0",
-    "async-timeout>=4.0.0",
-    "certifi>=2024.2.2"
+    "aiofiles>=23.0.0,<25.0.0",
+    "google-genai>=1.16.0,<2.0.0"
   ],
   "single_config_entry": false,
-  "ssdp": [],
-  "usb": [],
-  "version": "2.2.0",
-  "zeroconf": []
+  "version": "2.5.1"
 }

custom_components/ha_text_ai/metrics.py

@@ -0,0 +1,176 @@
+"""
+Metrics management for HA Text AI integration.
+
+@license: MIT (https://opensource.org/licenses/MIT)
+@author: SMKRV
+@github: https://github.com/smkrv/ha-text-ai
+@source: https://github.com/smkrv/ha-text-ai
+"""
+from __future__ import annotations
+
+import json
+import logging
+import os
+import re
+import traceback
+from typing import Any
+
+from homeassistant.core import HomeAssistant
+from homeassistant.exceptions import HomeAssistantError
+from homeassistant.util import dt as dt_util
+
+_LOGGER = logging.getLogger(__name__)
+
+DEFAULT_METRICS: dict[str, Any] = {
+    "total_tokens": 0,
+    "prompt_tokens": 0,
+    "completion_tokens": 0,
+    "successful_requests": 0,
+    "failed_requests": 0,
+    "total_errors": 0,
+    "average_latency": 0,
+    "max_latency": 0,
+    "min_latency": 0,
+}
+
+class MetricsManager:
+    """Manages performance metrics for an instance."""
+
+    def __init__(
+        self,
+        hass: HomeAssistant,
+        instance_name: str,
+        metrics_file: str,
+    ) -> None:
+        self.hass = hass
+        self.instance_name = instance_name
+        self._metrics_file = metrics_file
+        self._performance_metrics: dict[str, Any] = DEFAULT_METRICS.copy()
+
+    @property
+    def metrics(self) -> dict[str, Any]:
+        return self._performance_metrics
+
+    async def async_initialize(self) -> None:
+        """Load metrics from storage or create defaults."""
+        loaded = await self._load_metrics()
+        self._performance_metrics = loaded or DEFAULT_METRICS.copy()
+
+    async def _load_metrics(self) -> dict[str, Any] | None:
+        try:
+            exists = await self.hass.async_add_executor_job(
+                os.path.exists, self._metrics_file
+            )
+            if exists:
+                def read_metrics():
+                    with open(self._metrics_file, "r") as f:
+                        try:
+                            return json.load(f)
+                        except json.JSONDecodeError:
+                            _LOGGER.warning("Metrics file corrupted, creating new")
+                            return None
+
+                return await self.hass.async_add_executor_job(read_metrics)
+        except Exception as e:
+            _LOGGER.warning("Failed to load metrics: %s", e)
+        return None
+
+    async def _save_metrics(self) -> None:
+        try:
+            def write_metrics():
+                with open(self._metrics_file, "w") as f:
+                    json.dump(self._performance_metrics, f)
+
+            await self.hass.async_add_executor_job(write_metrics)
+        except Exception as e:
+            _LOGGER.warning("Failed to save metrics: %s", e)
+
+    async def update_metrics(self, latency: float, response: dict) -> None:
+        """Update performance metrics after a successful request."""
+        metrics = self._performance_metrics
+        tokens = response.get("tokens", {})
+
+        metrics["total_tokens"] += tokens.get("total", 0)
+        metrics["prompt_tokens"] += tokens.get("prompt", 0)
+        metrics["completion_tokens"] += tokens.get("completion", 0)
+        metrics["successful_requests"] += 1
+
+        metrics["average_latency"] = (
+            (metrics["average_latency"] * (metrics["successful_requests"] - 1) + latency)
+            / metrics["successful_requests"]
+        )
+        metrics["max_latency"] = max(metrics["max_latency"], latency)
+        if metrics["min_latency"] == 0:
+            metrics["min_latency"] = latency
+        else:
+            metrics["min_latency"] = min(metrics["min_latency"], latency)
+
+        await self._save_metrics()
+
+    async def get_current_metrics(self) -> dict[str, Any]:
+        """Get current performance metrics."""
+        return self._performance_metrics.copy()
+
+    async def handle_error(
+        self,
+        error: Exception,
+        model: str,
+    ) -> dict[str, Any]:
+        """Record an error in metrics and return error details."""
+        self._performance_metrics["total_errors"] += 1
+        self._performance_metrics["failed_requests"] += 1
+        await self._save_metrics()
+
+        error_msg = str(error)
+        # Strip URLs, API keys, tokens, and query parameters from error messages.
+        # Patterns use word boundaries and explicit length bounds so that
+        # overly greedy matches don't accidentally swallow adjacent text.
+        error_msg = re.sub(r'https?://\S+', '[URL]', error_msg)
+        error_msg = re.sub(r'[?&]key=[^\s&]+', '?key=***', error_msg)
+        # Google API key: fixed prefix + 30+ url-safe chars, bounded by non-key char.
+        error_msg = re.sub(
+            r'AIza[A-Za-z0-9_\-]{30,}(?=[^A-Za-z0-9_\-]|$)', '***', error_msg
+        )
+        # Anthropic / OpenAI / DeepSeek format: "sk-..." (anchors on word boundary).
+        error_msg = re.sub(r'\bsk-[A-Za-z0-9_\-]{20,}\b', '***', error_msg)
+        # Bearer tokens: header-style and JSON-embedded ("Bearer xxx").
+        error_msg = re.sub(r'[Bb]earer\s+[A-Za-z0-9_\-\.=]+', 'Bearer ***', error_msg)
+        # x-api-key header in any case, both raw and JSON-serialized forms.
+        error_msg = re.sub(
+            r'"?x-api-key"?\s*[:=]\s*"?[A-Za-z0-9_\-\.]+"?',
+            'x-api-key: ***',
+            error_msg,
+            flags=re.IGNORECASE,
+        )
+        if len(error_msg) > 256:
+            error_msg = error_msg[:256] + "..."
+
+        error_details: dict[str, Any] = {
+            "timestamp": dt_util.utcnow().isoformat(),
+            "model": model,
+            "instance": self.instance_name,
+            "error_message": error_msg,
+            "error_type": type(error).__name__,
+            "traceback": traceback.format_exc()
+            if _LOGGER.isEnabledFor(logging.DEBUG)
+            else None,
+        }
+
+        error_mapping = {
+            HomeAssistantError: {"is_ha_error": True},
+            ConnectionError: {"is_connection_error": True},
+            TimeoutError: {"is_timeout": True},
+            PermissionError: {"is_permission_denied": True},
+            ValueError: {"is_validation_error": True},
+        }
+
+        for error_type, error_flags in error_mapping.items():
+            if isinstance(error, error_type):
+                error_details.update(error_flags)
+                break
+
+        _LOGGER.error("AI Processing Error: %s", error_details)
+        if _LOGGER.isEnabledFor(logging.DEBUG):
+            _LOGGER.debug("Full Error Traceback: %s", error_details.get("traceback"))
+
+        return error_details

custom_components/ha_text_ai/providers.py

@@ -0,0 +1,93 @@
+"""
+Provider registry for HA Text AI integration.
+
+Centralizes provider-specific configuration to avoid dispatch duplication
+across __init__.py, config_flow.py, and api_client.py.
+
+@license: MIT (https://opensource.org/licenses/MIT)
+@author: SMKRV
+@github: https://github.com/smkrv/ha-text-ai
+@source: https://github.com/smkrv/ha-text-ai
+"""
+from __future__ import annotations
+
+from typing import Any
+
+from .const import (
+    API_PROVIDER_OPENAI,
+    API_PROVIDER_ANTHROPIC,
+    API_PROVIDER_DEEPSEEK,
+    API_PROVIDER_GEMINI,
+    DEFAULT_MODEL,
+    DEFAULT_ANTHROPIC_MODEL,
+    DEFAULT_DEEPSEEK_MODEL,
+    DEFAULT_GEMINI_MODEL,
+    DEFAULT_OPENAI_ENDPOINT,
+    DEFAULT_ANTHROPIC_ENDPOINT,
+    DEFAULT_DEEPSEEK_ENDPOINT,
+    DEFAULT_GEMINI_ENDPOINT,
+)
+
+PROVIDER_REGISTRY: dict[str, dict[str, Any]] = {
+    API_PROVIDER_OPENAI: {
+        "default_model": DEFAULT_MODEL,
+        "default_endpoint": DEFAULT_OPENAI_ENDPOINT,
+        "auth_header": "Authorization",
+        "auth_prefix": "Bearer ",
+        "check_path": "/models",
+    },
+    API_PROVIDER_ANTHROPIC: {
+        "default_model": DEFAULT_ANTHROPIC_MODEL,
+        "default_endpoint": DEFAULT_ANTHROPIC_ENDPOINT,
+        "auth_header": "x-api-key",
+        "auth_prefix": "",
+        "check_path": "/v1/models",
+        "extra_headers": {
+            "anthropic-version": "2023-06-01",
+        },
+    },
+    API_PROVIDER_DEEPSEEK: {
+        "default_model": DEFAULT_DEEPSEEK_MODEL,
+        "default_endpoint": DEFAULT_DEEPSEEK_ENDPOINT,
+        "auth_header": "Authorization",
+        "auth_prefix": "Bearer ",
+        "check_path": "/models",
+    },
+    API_PROVIDER_GEMINI: {
+        "default_model": DEFAULT_GEMINI_MODEL,
+        "default_endpoint": DEFAULT_GEMINI_ENDPOINT,
+        "auth_header": "Authorization",
+        "auth_prefix": "Bearer ",
+        "check_path": None,  # Gemini does not support /models check
+    },
+}
+
+def get_provider_config(provider: str) -> dict[str, Any]:
+    """Get full provider configuration.
+
+    Raises ValueError for unknown providers to avoid sending
+    credentials to the wrong endpoint.
+    """
+    if provider not in PROVIDER_REGISTRY:
+        raise ValueError(f"Unknown API provider: {provider}")
+    return PROVIDER_REGISTRY[provider]
+
+def get_default_endpoint(provider: str) -> str:
+    """Get default API endpoint for a provider."""
+    return get_provider_config(provider)["default_endpoint"]
+
+def get_default_model(provider: str) -> str:
+    """Get default model for a provider."""
+    return get_provider_config(provider)["default_model"]
+
+def build_auth_headers(provider: str, api_key: str) -> dict[str, str]:
+    """Build authentication headers for a provider."""
+    config = get_provider_config(provider)
+    headers = {
+        "Content-Type": "application/json",
+        "Accept": "application/json",
+    }
+    headers[config["auth_header"]] = f"{config['auth_prefix']}{api_key}"
+    if "extra_headers" in config:
+        headers.update(config["extra_headers"])
+    return headers

custom_components/ha_text_ai/sensor.py

@@ -1,23 +1,23 @@
 """
 Sensor platform for HA Text AI.
 
-@license: CC BY-NC-SA 4.0 International
+@license: MIT (https://opensource.org/licenses/MIT)
 @author: SMKRV
 @github: https://github.com/smkrv/ha-text-ai
 @source: https://github.com/smkrv/ha-text-ai
 """
+from __future__ import annotations
+
 import logging
 import math
-from typing import Any, Dict
-from datetime import datetime, timedelta
-
+from typing import Any
 from homeassistant.components.sensor import (
     SensorEntity,
     SensorEntityDescription,
 )
 from homeassistant.config_entries import ConfigEntry
 from homeassistant.core import HomeAssistant
-from homeassistant.helpers.device_registry import DeviceInfo
+from homeassistant.helpers.device_registry import DeviceEntryType, DeviceInfo
 from homeassistant.helpers.entity_platform import AddEntitiesCallback
 from homeassistant.helpers.typing import StateType
 from homeassistant.helpers.update_coordinator import CoordinatorEntity
@@ -44,10 +44,8 @@ from .const import (
     ATTR_API_PROVIDER,
     ATTR_MODEL,
     ATTR_SYSTEM_PROMPT,
-    ATTR_API_STATUS,
     ATTR_RESPONSE,
     ATTR_QUESTION,
-    ATTR_CONVERSATION_HISTORY,
     METRIC_TOTAL_TOKENS,
     METRIC_PROMPT_TOKENS,
     METRIC_COMPLETION_TOKENS,
@@ -68,14 +66,18 @@ from .const import (
     ENTITY_ICON_PROCESSING,
     DEFAULT_NAME_PREFIX,
     CONF_MAX_HISTORY_SIZE,
-    MAX_ATTRIBUTE_SIZE,
     VERSION,
 )
 
 from .coordinator import HATextAICoordinator
+from .utils import safe_log_data
 
 _LOGGER = logging.getLogger(__name__)
 
+# HA Recorder limit is 16384 bytes for state_attributes.
+# Budget per field to stay well within the limit.
+_ATTR_TEXT_LIMIT = 2048
+_ATTR_PROMPT_LIMIT = 512
 
 async def async_setup_entry(
     hass: HomeAssistant,
@@ -83,23 +85,23 @@ async def async_setup_entry(
     async_add_entities: AddEntitiesCallback,
 ) -> None:
     """Set up the HA Text AI sensor."""
-    _LOGGER.debug(f"Starting sensor setup for entry: {entry.entry_id}")
+    _LOGGER.debug("Starting sensor setup for entry: %s", entry.entry_id)
 
     try:
         coordinator = hass.data[DOMAIN][entry.entry_id]
-        _LOGGER.debug(f"Found coordinator for entry {entry.entry_id}")
+        _LOGGER.debug("Found coordinator for entry %s", entry.entry_id)
 
         instance_name = coordinator.instance_name
-        _LOGGER.debug(f"Setting up sensor with instance: {instance_name}")
+        _LOGGER.debug("Setting up sensor with instance: %s", instance_name)
 
         sensor = HATextAISensor(coordinator, entry)
-        _LOGGER.debug(f"Created sensor instance: {sensor.entity_id}")
+        _LOGGER.debug("Created sensor instance: %s", sensor.entity_id)
 
         async_add_entities([sensor], True)
-        _LOGGER.debug(f"Added sensor entity: {sensor.entity_id}")
+        _LOGGER.debug("Added sensor entity: %s", sensor.entity_id)
 
     except Exception as err:
-        _LOGGER.exception(f"Error setting up sensor: {err}")
+        _LOGGER.exception("Error setting up sensor: %s", err)
         raise
 
 class HATextAISensor(CoordinatorEntity, SensorEntity):
@@ -113,7 +115,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
         config_entry: ConfigEntry,
     ) -> None:
         """Initialize the sensor."""
-        _LOGGER.debug(f"Initializing sensor with config entry: {config_entry.data}")
+        _LOGGER.debug("Initializing sensor with config entry: %s", safe_log_data(dict(config_entry.data)))
 
         super().__init__(coordinator)
 
@@ -121,19 +123,20 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
         self._instance_name = coordinator.instance_name
         self._normalized_name = coordinator.normalized_name
 
-        _LOGGER.debug(f"Instance name: {self._instance_name}")
-        _LOGGER.debug(f"Normalized name: {self._normalized_name}")
+        _LOGGER.debug("Instance name: %s", self._instance_name)
+        _LOGGER.debug("Normalized name: %s", self._normalized_name)
 
         self._conversation_history = []
         self._system_prompt = None
 
-        self._attr_name = f"HA Text AI {self._instance_name}"
+        self._attr_has_entity_name = True
+        self._attr_name = self._instance_name
         self.entity_id = f"sensor.ha_text_ai_{self._normalized_name}"
-        self._attr_unique_id = f"{config_entry.entry_id}"
+        self._attr_unique_id = config_entry.entry_id
 
-        _LOGGER.debug(f"Created sensor with entity_id: {self.entity_id}")
-        _LOGGER.debug(f"Sensor name: {self._attr_name}")
-        _LOGGER.debug(f"Unique ID: {self._attr_unique_id}")
+        _LOGGER.debug("Created sensor with entity_id: %s", self.entity_id)
+        _LOGGER.debug("Sensor name: %s", self._attr_name)
+        _LOGGER.debug("Unique ID: %s", self._attr_unique_id)
 
         self.entity_description = SensorEntityDescription(
             key=f"ha_text_ai_{self._normalized_name.lower()}",
@@ -157,10 +160,12 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
             manufacturer="Community",
             model=f"{model} ({api_provider} provider)",
             sw_version=VERSION,
+            entry_type=DeviceEntryType.SERVICE,
         )
 
         _LOGGER.debug(
-            f"Initialized sensor: {self.entity_id} for instance: {self._instance_name}"
+            "Initialized sensor: %s for instance: %s",
+            self.entity_id, self._instance_name,
         )
 
     @property
@@ -179,7 +184,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
                 return None
         return value
 
-    def _sanitize_attributes(self, attributes: Dict[str, Any]) -> Dict[str, Any]:
+    def _sanitize_attributes(self, attributes: dict[str, Any]) -> dict[str, Any]:
         """Sanitize all attributes for JSON serialization."""
         sanitized = {
             key: self._sanitize_value(value)
@@ -200,7 +205,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
         ]
 
         metrics_values = {k: sanitized.get(k) for k in metrics_keys if k in sanitized}
-        _LOGGER.debug(f"Metrics for {self.entity_id}: {metrics_values}")
+        _LOGGER.debug("Metrics for %s: %s", self.entity_id, metrics_values)
 
         return sanitized
 
@@ -225,7 +230,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
         return ENTITY_ICON
 
     @property
-    def extra_state_attributes(self) -> Dict[str, Any]:
+    def extra_state_attributes(self) -> dict[str, Any]:
         """Return entity specific state attributes."""
         if not self.coordinator.data:
             return {}
@@ -238,11 +243,10 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
             attributes = {
                 ATTR_MODEL: self._config_entry.data.get(CONF_MODEL, "Unknown"),
                 ATTR_API_PROVIDER: self._config_entry.data.get(CONF_API_PROVIDER, "Unknown"),
-                ATTR_API_STATUS: self._current_state,
                 ATTR_TOTAL_ERRORS: metrics.get("total_errors", 0),
                 "instance_name": self._instance_name,
                 "normalized_name": self._normalized_name,
-                ATTR_SYSTEM_PROMPT: (data.get("system_prompt", "")[:MAX_ATTRIBUTE_SIZE]
+                ATTR_SYSTEM_PROMPT: (data.get("system_prompt", "")[:_ATTR_PROMPT_LIMIT]
                                    if data.get("system_prompt") else None),
                 ATTR_IS_PROCESSING: data.get("is_processing", False),
                 ATTR_IS_RATE_LIMITED: data.get("is_rate_limited", False),
@@ -252,18 +256,19 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
                 ATTR_HISTORY_SIZE: data.get("history_size", 0),
             }
 
-            # History limit
+            # Conversation history preview (compact: last 3, truncated to 256 chars).
+            # Full history is available via ha_text_ai.get_history service.
             conversation_history = data.get("conversation_history", [])
             if conversation_history:
-                limited_history = []
-                for entry in conversation_history:
-                    limited_entry = {
+                preview = conversation_history[-3:]
+                attributes["conversation_history"] = [
+                    {
                         "timestamp": entry["timestamp"],
-                        "question": entry["question"][:MAX_ATTRIBUTE_SIZE],
-                        "response": entry["response"][:MAX_ATTRIBUTE_SIZE]
+                        "question": entry["question"][:256],
+                        "response": entry["response"][:256],
                     }
-                    limited_history.append(limited_entry)
-                attributes[ATTR_CONVERSATION_HISTORY] = limited_history
+                    for entry in preview
+                ]
 
             # Metrics
             if isinstance(metrics, dict):
@@ -275,20 +280,18 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
                     METRIC_FAILED_REQUESTS: metrics.get("failed_requests", 0),
                     METRIC_AVERAGE_LATENCY: round(metrics.get("average_latency", 0), 2),
                     METRIC_MAX_LATENCY: round(metrics.get("max_latency", 0), 2),
-                    METRIC_MIN_LATENCY: (metrics.get("min_latency")
-                                       if metrics.get("min_latency") != float("inf")
-                                       else None),
+                    METRIC_MIN_LATENCY: metrics.get("min_latency", 0) or None,
                 })
 
             # Last response handling
             last_response = data.get("last_response", {})
             if isinstance(last_response, dict):
                 attributes.update({
-                    ATTR_RESPONSE: last_response.get("response", "")[:MAX_ATTRIBUTE_SIZE],
-                    ATTR_QUESTION: last_response.get("question", "")[:MAX_ATTRIBUTE_SIZE],
+                    ATTR_RESPONSE: last_response.get("response", "")[:_ATTR_TEXT_LIMIT],
+                    ATTR_QUESTION: last_response.get("question", "")[:_ATTR_TEXT_LIMIT],
                     "last_model": last_response.get("model", ""),
                     "last_timestamp": last_response.get("timestamp", ""),
-                    "last_error": (last_response.get("error", "")[:MAX_ATTRIBUTE_SIZE]
+                    "last_error": (last_response.get("error", "")[:_ATTR_TEXT_LIMIT]
                                  if last_response.get("error") else None),
                 })
 
@@ -302,7 +305,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
         """When entity is added to hass."""
         await super().async_added_to_hass()
         self._handle_coordinator_update()
-        _LOGGER.debug(f"Entity {self.entity_id} added to Home Assistant")
+        _LOGGER.debug("Entity %s added to Home Assistant", self.entity_id)
 
     def _handle_coordinator_update(self) -> None:
         """Handle updated data from the coordinator."""
@@ -310,7 +313,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
             data = self.coordinator.data
             if not self.coordinator.last_update_success or not data:
                 self._current_state = STATE_DISCONNECTED
-                _LOGGER.warning(f"No data available for {self.entity_id}")
+                _LOGGER.warning("No data available for %s", self.entity_id)
                 self.async_write_ha_state()
                 return
 
@@ -320,7 +323,7 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
             metrics = data.get("metrics", {})
             if isinstance(metrics, dict):
                 self._metrics.update(metrics)
-                _LOGGER.debug(f"Updated metrics for {self.entity_id}: {self._metrics}")
+                _LOGGER.debug("Updated metrics for %s: %s", self.entity_id, self._metrics)
 
             # Update conversation history and system prompt
             self._conversation_history = data.get("conversation_history", [])
@@ -344,8 +347,8 @@ class HATextAISensor(CoordinatorEntity, SensorEntity):
             self._last_update = dt_util.utcnow()
 
             _LOGGER.debug(
-                f"Updated {self.entity_id} state to: {self._current_state} "
-                f"(available: {self.available})"
+                "Updated %s state to: %s (available: %s)",
+                self.entity_id, self._current_state, self.available,
             )
 
         except Exception as err:

custom_components/ha_text_ai/services.yaml

@@ -74,6 +74,33 @@ ask_question:
           step: 1
           mode: box
 
+    structured_output:
+      name: Structured Output
+      description: Enable JSON structured output mode. When enabled, the AI will respond with valid JSON matching the provided schema.
+      required: false
+      default: false
+      selector:
+        boolean: {}
+
+    json_schema:
+      name: JSON Schema
+      description: >-
+        JSON Schema defining the structure of the expected response.
+        Required when structured_output is enabled.
+      required: false
+      selector:
+        text:
+          multiline: true
+
+    disable_thinking:
+      name: Disable Thinking
+      description: >-
+        Disable model thinking/reasoning for this request.
+        Overrides the integration-level setting when provided.
+      required: false
+      selector:
+        boolean: {}
+
 clear_history:
   name: Clear History
   description: >-
@@ -135,7 +162,7 @@ get_history:
       required: false
       default: false
       selector:
-        boolean:
+        boolean: {}
 
     sort_order:
       name: Sort Order

custom_components/ha_text_ai/strings.json

@@ -0,0 +1,336 @@
+{
+  "config": {
+    "step": {
+      "provider": {
+        "title": "Provider Settings",
+        "description": "Provide connection details for your chosen AI provider.",
+        "data": {
+          "name": "Instance name (e.g., 'GPT Assistant', 'Claude Helper')",
+          "api_key": "API key for authentication",
+          "model": "AI model to use",
+          "api_endpoint": "Custom API endpoint URL (optional)",
+          "temperature": "Response creativity (0-2, lower = more focused)",
+          "max_tokens": "Maximum response length (1-100000 tokens)",
+          "request_interval": "Minimum time between requests (0.1-60 seconds)",
+          "api_timeout": "API request timeout in seconds (5-600)",
+          "context_messages": "Number of context messages to retain (1-20)",
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
+        }
+      },
+      "user": {
+        "title": "Configure HA Text AI Instance",
+        "description": "Set up a new AI assistant instance with your selected provider.",
+        "data": {
+          "name": "Instance name (e.g., 'GPT Assistant', 'Claude Helper')",
+          "api_key": "API key for authentication",
+          "model": "AI model to use",
+          "temperature": "Response creativity (0-2, lower = more focused)",
+          "max_tokens": "Maximum response length (1-100000 tokens)",
+          "api_endpoint": "Custom API endpoint URL (optional)",
+          "api_provider": "API Provider",
+          "request_interval": "Minimum time between requests (0.1-60 seconds)",
+          "api_timeout": "API request timeout in seconds (5-600)",
+          "context_messages": "Number of context messages to retain (1-20)",
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
+        }
+      }
+    },
+    "error": {
+      "history_storage_error": "Failed to initialize history storage. Check permissions.",
+      "history_rotation_error": "Error during history file rotation.",
+      "history_file_access_error": "Cannot access history storage directory.",
+      "name_exists": "An instance with this name already exists",
+      "invalid_name": "Invalid instance name",
+      "invalid_auth": "Authentication failed - check your API key",
+      "api_key_required": "API key is required when changing provider or endpoint",
+      "invalid_api_key": "Invalid API key - please verify your credentials",
+      "cannot_connect": "Failed to connect to API service",
+      "invalid_model": "Selected model is not available",
+      "rate_limit": "Rate limit exceeded",
+      "context_length": "Context length exceeded",
+      "rate_limit_exceeded": "API rate limit exceeded",
+      "maintenance": "Service is under maintenance",
+      "invalid_response": "Invalid API response received",
+      "api_error": "API service error occurred",
+      "timeout": "Request timed out",
+      "invalid_instance": "Invalid instance specified",
+      "unknown": "Unexpected error occurred",
+      "empty": "Name cannot be empty",
+      "name_too_long": "Name must be 50 characters or less"
+    },
+    "abort": {
+      "already_configured": "Instance already configured"
+    }
+  },
+  "options": {
+    "step": {
+      "init": {
+        "title": "Select Provider",
+        "description": "Choose the AI provider for this instance. The integration will reload after saving changes.",
+        "data": {
+          "api_provider": "API Provider"
+        }
+      },
+      "settings": {
+        "title": "Connection & Model Settings",
+        "description": "Configure API credentials and model parameters. Changes will take effect after the integration reloads.",
+        "data": {
+          "api_key": "API Key",
+          "api_endpoint": "API Endpoint URL",
+          "model": "AI model",
+          "temperature": "Response creativity (0-2)",
+          "max_tokens": "Maximum response length (1-100000)",
+          "request_interval": "Minimum request interval (0.1-60 seconds)",
+          "api_timeout": "API request timeout in seconds (5-600)",
+          "context_messages": "Number of previous messages to include in context (1-20)",
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
+        }
+      }
+    }
+  },
+  "selector": {
+    "api_provider": {
+      "options": {
+        "openai": "OpenAI (compatible)",
+        "anthropic": "Anthropic (compatible)",
+        "deepseek": "DeepSeek",
+        "gemini": "Google Gemini"
+      }
+    }
+  },
+  "services": {
+    "ask_question": {
+      "name": "Ask Question (HA Text AI)",
+      "description": "Send a question to the AI model and receive a detailed response. This service now returns response data directly, eliminating the need for separate text sensors and the 255-character limitation. The response will also be stored in the conversation history.",
+      "fields": {
+        "instance": {
+          "name": "Instance",
+          "description": "Name of the HA Text AI instance to use"
+        },
+        "question": {
+          "name": "Question",
+          "description": "Your question or prompt for the AI assistant"
+        },
+        "context_messages": {
+          "name": "Context Messages",
+          "description": "Number of previous messages to include in context (1-20)"
+        },
+        "system_prompt": {
+          "name": "System Prompt",
+          "description": "Optional system prompt to set context for this specific question"
+        },
+        "model": {
+          "name": "Model",
+          "description": "Select AI model to use (optional, overrides default setting)"
+        },
+        "temperature": {
+          "name": "Temperature",
+          "description": "Controls response creativity (0.0-2.0)"
+        },
+        "max_tokens": {
+          "name": "Max Tokens",
+          "description": "Maximum length of the response (1-100000 tokens)"
+        },
+        "structured_output": {
+          "name": "Structured Output",
+          "description": "Enable JSON structured output mode. When enabled, the AI will respond with valid JSON matching the provided schema."
+        },
+        "json_schema": {
+          "name": "JSON Schema",
+          "description": "JSON Schema defining the structure of the expected response. Required when structured_output is enabled."
+        },
+        "disable_thinking": {
+          "name": "Disable Thinking",
+          "description": "Disable model thinking/reasoning for this request. Overrides the integration-level setting."
+        }
+      }
+    },
+    "clear_history": {
+      "name": "Clear History",
+      "description": "Delete all stored questions and responses from the conversation history",
+      "fields": {
+        "instance": {
+          "name": "Instance",
+          "description": "Name of the HA Text AI instance to clear history for"
+        }
+      }
+    },
+    "get_history": {
+      "name": "Get History",
+      "description": "Retrieve conversation history with optional filtering and sorting",
+      "fields": {
+        "instance": {
+          "name": "Instance",
+          "description": "Name of the HA Text AI instance to get history from"
+        },
+        "limit": {
+          "name": "Limit",
+          "description": "Number of conversations to return (1-100)"
+        },
+        "filter_model": {
+          "name": "Filter Model",
+          "description": "Filter conversations by specific AI model"
+        },
+        "start_date": {
+          "name": "Start Date",
+          "description": "Filter conversations starting from this date/time"
+        },
+        "include_metadata": {
+          "name": "Include Metadata",
+          "description": "Include additional information like tokens used, response time, etc."
+        },
+        "sort_order": {
+          "name": "Sort Order",
+          "description": "Sort order for results (newest or oldest first)"
+        }
+      }
+    },
+    "set_system_prompt": {
+      "name": "Set System Prompt",
+      "description": "Set default system behavior instructions for all future conversations",
+      "fields": {
+        "instance": {
+          "name": "Instance",
+          "description": "Name of the HA Text AI instance to set system prompt for"
+        },
+        "prompt": {
+          "name": "System Prompt",
+          "description": "Instructions that define how the AI should behave and respond"
+        }
+      }
+    }
+  },
+  "entity": {
+    "sensor": {
+      "ha_text_ai": {
+        "name": "{name}",
+        "state": {
+          "ready": "Ready",
+          "processing": "Processing",
+          "error": "Error",
+          "disconnected": "Disconnected",
+          "rate_limited": "Rate Limited",
+          "maintenance": "Maintenance",
+          "initializing": "Initializing",
+          "retrying": "Retrying"
+        },
+        "state_attributes": {
+          "question": {
+            "name": "Last Question"
+          },
+          "response": {
+            "name": "Last Response"
+          },
+          "model": {
+            "name": "Current Model"
+          },
+          "temperature": {
+            "name": "Temperature"
+          },
+          "max_tokens": {
+            "name": "Max Tokens"
+          },
+          "system_prompt": {
+            "name": "System Prompt"
+          },
+          "response_time": {
+            "name": "Last Response Time"
+          },
+          "total_responses": {
+            "name": "Total Responses"
+          },
+          "error_count": {
+            "name": "Error Count"
+          },
+          "last_error": {
+            "name": "Last Error"
+          },
+          "api_status": {
+            "name": "API Status"
+          },
+          "tokens_used": {
+            "name": "Total Tokens Used"
+          },
+          "average_response_time": {
+            "name": "Average Response Time"
+          },
+          "last_request_time": {
+            "name": "Last Request Time"
+          },
+          "is_processing": {
+            "name": "Processing Status"
+          },
+          "is_rate_limited": {
+            "name": "Rate Limited Status"
+          },
+          "is_maintenance": {
+            "name": "Maintenance Status"
+          },
+          "api_version": {
+            "name": "API Version"
+          },
+          "endpoint_status": {
+            "name": "Endpoint Status"
+          },
+          "performance_metrics": {
+            "name": "Performance Metrics"
+          },
+          "history_size": {
+            "name": "History Size"
+          },
+          "uptime": {
+            "name": "Uptime"
+          },
+          "total_tokens": {
+            "name": "Total Tokens"
+          },
+          "prompt_tokens": {
+            "name": "Prompt Tokens"
+          },
+          "completion_tokens": {
+            "name": "Completion Tokens"
+          },
+          "successful_requests": {
+            "name": "Successful Requests"
+          },
+          "failed_requests": {
+            "name": "Failed Requests"
+          },
+          "average_latency": {
+            "name": "Average Latency"
+          },
+          "max_latency": {
+            "name": "Maximum Latency"
+          },
+          "min_latency": {
+            "name": "Minimum Latency"
+          },
+          "last_model": {
+            "name": "Last Used Model"
+          },
+          "last_timestamp": {
+            "name": "Last Response Time"
+          },
+          "instance_name": {
+            "name": "Instance Name"
+          },
+          "normalized_name": {
+            "name": "Normalized Name"
+          },
+          "last_error": {
+            "name": "Last Error"
+          },
+          "conversation_history": {
+            "name": "Conversation History"
+          }
+        }
+      }
+    }
+  }
+}

custom_components/ha_text_ai/translations/de.json

@@ -1,15 +1,6 @@
 {
   "config": {
     "step": {
-      "provider": {
-        "title": "Wählen Sie AI-Anbieter",
-        "description": "Wählen Sie, welchen AI-Dienstanbieter Sie für diese Instanz verwenden möchten.",
-        "data": {
-          "api_provider": "API-Anbieter",
-          "context_messages": "Anzahl der zu behaltenden Kontextnachrichten (1-20)",
-          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)"
-        }
-      },
       "provider": {
         "title": "Anbieter-Einstellungen",
         "description": "Geben Sie die Verbindungsdetails für Ihren gewählten AI-Anbieter an.",
@@ -23,7 +14,9 @@
           "request_interval": "Minimale Zeit zwischen Anfragen (0,1-60 Sekunden)",
           "api_timeout": "API-Anfrage Timeout in Sekunden (5-600)",
           "context_messages": "Anzahl der zu behaltenden Kontextnachrichten (1-20)",
-          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)"
+          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)",
+          "allow_local_network": "Lokale Netzwerkendpunkte erlauben (für selbst gehostete Proxys)",
+          "disable_thinking": "Thinking/Reasoning-Modus deaktivieren (Qwen /no_think, think-Blöcke entfernen, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -40,7 +33,9 @@
           "request_interval": "Minimale Zeit zwischen Anfragen (0,1-60 Sekunden)",
           "api_timeout": "API-Anfrage Timeout in Sekunden (5-600)",
           "context_messages": "Anzahl der zu behaltenden Kontextnachrichten (1-20)",
-          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)"
+          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)",
+          "allow_local_network": "Lokale Netzwerkendpunkte erlauben (für selbst gehostete Proxys)",
+          "disable_thinking": "Thinking/Reasoning-Modus deaktivieren (Qwen /no_think, think-Blöcke entfernen, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -51,6 +46,7 @@
       "name_exists": "Eine Instanz mit diesem Namen existiert bereits",
       "invalid_name": "Ungültiger Instanzname",
       "invalid_auth": "Authentifizierung fehlgeschlagen - überprüfen Sie Ihren API-Schlüssel",
+      "api_key_required": "API-Schlüssel ist erforderlich, wenn Anbieter oder Endpunkt geändert wird",
       "invalid_api_key": "Ungültiger API-Schlüssel - bitte überprüfen Sie Ihre Anmeldeinformationen",
       "cannot_connect": "Verbindung zum API-Dienst fehlgeschlagen",
       "invalid_model": "Ausgewähltes Modell ist nicht verfügbar",
@@ -64,7 +60,6 @@
       "invalid_instance": "Ungültige Instanz angegeben",
       "unknown": "Unerwarteter Fehler aufgetreten",
       "empty": "Name darf nicht leer sein",
-      "invalid_characters": "Name darf nur Buchstaben, Zahlen, Leerzeichen, Unterstriche und Bindestriche enthalten",
       "name_too_long": "Name darf höchstens 50 Zeichen lang sein"
     },
     "abort": {
@@ -74,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Instanzeinstellungen aktualisieren",
-        "description": "Ändern Sie die Einstellungen für diese AI-Assistenteninstanz.",
+        "title": "Anbieter auswählen",
+        "description": "Wählen Sie den AI-Anbieter für diese Instanz. Die Integration wird nach dem Speichern der Änderungen neu geladen.",
         "data": {
+          "api_provider": "API-Anbieter"
+        }
+      },
+      "settings": {
+        "title": "Verbindungs- und Modelleinstellungen",
+        "description": "Konfigurieren Sie API-Anmeldeinformationen und Modellparameter. Änderungen werden nach dem Neuladen der Integration wirksam.",
+        "data": {
+          "api_key": "API-Schlüssel",
+          "api_endpoint": "API-Endpunkt-URL",
           "model": "AI-Modell",
           "temperature": "Kreativität der Antwort (0-2)",
           "max_tokens": "Maximale Länge der Antwort (1-100000)",
-          "request_interval": "Minimale Anfrageintervall (0,1-60 Sekunden)",
+          "request_interval": "Minimales Anfrageintervall (0,1-60 Sekunden)",
           "api_timeout": "API-Anfrage Timeout in Sekunden (5-600)",
           "context_messages": "Anzahl der vorherigen Nachrichten, die im Kontext enthalten sein sollen (1-20)",
-          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)"
+          "max_history_size": "Maximale Größe des Gesprächsverlaufs (1-100)",
+          "allow_local_network": "Lokale Netzwerkendpunkte erlauben (für selbst gehostete Proxys)",
+          "disable_thinking": "Thinking/Reasoning-Modus deaktivieren (Qwen /no_think, think-Blöcke entfernen, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -130,6 +136,18 @@
         "max_tokens": {
           "name": "Max Tokens",
           "description": "Maximale Länge der Antwort (1-100000 Token)"
+        },
+        "structured_output": {
+          "name": "Strukturierte Ausgabe",
+          "description": "JSON-Strukturausgabemodus aktivieren. Bei Aktivierung antwortet die KI mit gültigem JSON, das dem angegebenen Schema entspricht."
+        },
+        "json_schema": {
+          "name": "JSON-Schema",
+          "description": "JSON-Schema, das die Struktur der erwarteten Antwort definiert. Erforderlich wenn structured_output aktiviert ist."
+        },
+        "disable_thinking": {
+          "name": "Thinking deaktivieren",
+          "description": "Thinking/Reasoning-Modus für diese Anfrage deaktivieren. Überschreibt die Integrationseinstellung."
         }
       }
     },
@@ -200,8 +218,7 @@
           "rate_limited": "Rate limitiert",
           "maintenance": "Wartung",
           "initializing": "Initialisierung",
-          "retrying": "Wiederholen",
-          "queued": "In der Warteschlange"
+          "retrying": "Wiederholen"
         },
         "state_attributes": {
           "question": {
@@ -293,6 +310,21 @@
           },
           "min_latency": {
             "name": "Minimale Latenz"
+          },
+          "last_model": {
+            "name": "Zuletzt verwendetes Modell"
+          },
+          "last_timestamp": {
+            "name": "Zeitpunkt der letzten Antwort"
+          },
+          "instance_name": {
+            "name": "Instanzname"
+          },
+          "normalized_name": {
+            "name": "Normalisierter Name"
+          },
+          "conversation_history": {
+            "name": "Konversationsverlauf"
           }
         }
       }

custom_components/ha_text_ai/translations/en.json

@@ -1,15 +1,6 @@
 {
   "config": {
     "step": {
-      "provider": {
-        "title": "Select AI Provider",
-        "description": "Choose which AI service provider to use for this instance.",
-        "data": {
-          "api_provider": "API Provider",
-          "context_messages": "Number of context messages to retain (1-20)",
-          "max_history_size": "Maximum conversation history size (1-100)"
-        }
-      },
       "provider": {
         "title": "Provider Settings",
         "description": "Provide connection details for your chosen AI provider.",
@@ -23,7 +14,9 @@
           "request_interval": "Minimum time between requests (0.1-60 seconds)",
           "api_timeout": "API request timeout in seconds (5-600)",
           "context_messages": "Number of context messages to retain (1-20)",
-          "max_history_size": "Maximum conversation history size (1-100)"
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -40,7 +33,9 @@
           "request_interval": "Minimum time between requests (0.1-60 seconds)",
           "api_timeout": "API request timeout in seconds (5-600)",
           "context_messages": "Number of context messages to retain (1-20)",
-          "max_history_size": "Maximum conversation history size (1-100)"
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -51,6 +46,7 @@
       "name_exists": "An instance with this name already exists",
       "invalid_name": "Invalid instance name",
       "invalid_auth": "Authentication failed - check your API key",
+      "api_key_required": "API key is required when changing provider or endpoint",
       "invalid_api_key": "Invalid API key - please verify your credentials",
       "cannot_connect": "Failed to connect to API service",
       "invalid_model": "Selected model is not available",
@@ -64,7 +60,6 @@
       "invalid_instance": "Invalid instance specified",
       "unknown": "Unexpected error occurred",
       "empty": "Name cannot be empty",
-      "invalid_characters": "Name can only contain letters, numbers, spaces, underscores and hyphens",
       "name_too_long": "Name must be 50 characters or less"
     },
     "abort": {
@@ -74,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Update Instance Settings",
-        "description": "Modify settings for this AI assistant instance.",
+        "title": "Select Provider",
+        "description": "Choose the AI provider for this instance. The integration will reload after saving changes.",
         "data": {
+          "api_provider": "API Provider"
+        }
+      },
+      "settings": {
+        "title": "Connection & Model Settings",
+        "description": "Configure API credentials and model parameters. Changes will take effect after the integration reloads.",
+        "data": {
+          "api_key": "API Key",
+          "api_endpoint": "API Endpoint URL",
           "model": "AI model",
           "temperature": "Response creativity (0-2)",
           "max_tokens": "Maximum response length (1-100000)",
           "request_interval": "Minimum request interval (0.1-60 seconds)",
           "api_timeout": "API request timeout in seconds (5-600)",
           "context_messages": "Number of previous messages to include in context (1-20)",
-          "max_history_size": "Maximum conversation history size (1-100)"
+          "max_history_size": "Maximum conversation history size (1-100)",
+          "allow_local_network": "Allow local network endpoints (for self-hosted proxies)",
+          "disable_thinking": "Disable thinking/reasoning mode (Qwen /no_think, strips think blocks, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -130,6 +136,18 @@
         "max_tokens": {
           "name": "Max Tokens",
           "description": "Maximum length of the response (1-100000 tokens)"
+        },
+        "structured_output": {
+          "name": "Structured Output",
+          "description": "Enable JSON structured output mode. When enabled, the AI will respond with valid JSON matching the provided schema."
+        },
+        "json_schema": {
+          "name": "JSON Schema",
+          "description": "JSON Schema defining the structure of the expected response. Required when structured_output is enabled."
+        },
+        "disable_thinking": {
+          "name": "Disable Thinking",
+          "description": "Disable model thinking/reasoning for this request. Overrides the integration-level setting."
         }
       }
     },
@@ -200,8 +218,7 @@
           "rate_limited": "Rate Limited",
           "maintenance": "Maintenance",
           "initializing": "Initializing",
-          "retrying": "Retrying",
-          "queued": "Queued"
+          "retrying": "Retrying"
         },
         "state_attributes": {
           "question": {
@@ -293,6 +310,24 @@
           },
           "min_latency": {
             "name": "Minimum Latency"
+          },
+          "last_model": {
+            "name": "Last Used Model"
+          },
+          "last_timestamp": {
+            "name": "Last Response Time"
+          },
+          "instance_name": {
+            "name": "Instance Name"
+          },
+          "normalized_name": {
+            "name": "Normalized Name"
+          },
+          "last_error": {
+            "name": "Last Error"
+          },
+          "conversation_history": {
+            "name": "Conversation History"
           }
         }
       }

custom_components/ha_text_ai/translations/es.json

@@ -1,15 +1,6 @@
 {
   "config": {
     "step": {
-      "provider": {
-        "title": "Seleccionar proveedor de IA",
-        "description": "Elige qué proveedor de servicio de IA utilizar para esta instancia.",
-        "data": {
-          "api_provider": "Proveedor de API",
-          "context_messages": "Número de mensajes de contexto a retener (1-20)",
-          "max_history_size": "Tamaño máximo del historial de conversación (1-100)"
-        }
-      },
       "provider": {
         "title": "Configuración del proveedor",
         "description": "Proporciona los detalles de conexión para tu proveedor de IA elegido.",
@@ -23,7 +14,9 @@
           "request_interval": "Tiempo mínimo entre solicitudes (0.1-60 segundos)",
           "api_timeout": "Tiempo de espera de solicitud API en segundos (5-600)",
           "context_messages": "Número de mensajes de contexto a retener (1-20)",
-          "max_history_size": "Tamaño máximo del historial de conversación (1-100)"
+          "max_history_size": "Tamaño máximo del historial de conversación (1-100)",
+          "allow_local_network": "Permitir endpoints de red local (para proxies autoalojados)",
+          "disable_thinking": "Desactivar modo thinking/reasoning (Qwen /no_think, elimina bloques think, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -40,7 +33,9 @@
           "request_interval": "Tiempo mínimo entre solicitudes (0.1-60 segundos)",
           "api_timeout": "Tiempo de espera de solicitud API en segundos (5-600)",
           "context_messages": "Número de mensajes de contexto a retener (1-20)",
-          "max_history_size": "Tamaño máximo del historial de conversación (1-100)"
+          "max_history_size": "Tamaño máximo del historial de conversación (1-100)",
+          "allow_local_network": "Permitir endpoints de red local (para proxies autoalojados)",
+          "disable_thinking": "Desactivar modo thinking/reasoning (Qwen /no_think, elimina bloques think, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -51,6 +46,7 @@
       "name_exists": "Ya existe una instancia con este nombre",
       "invalid_name": "Nombre de instancia no válido",
       "invalid_auth": "La autenticación falló - verifica tu clave API",
+      "api_key_required": "Se requiere la clave API al cambiar de proveedor o endpoint",
       "invalid_api_key": "Clave API no válida - verifica tus credenciales",
       "cannot_connect": "Error al conectar con el servicio de API",
       "invalid_model": "El modelo seleccionado no está disponible",
@@ -64,7 +60,6 @@
       "invalid_instance": "Instancia no válida especificada",
       "unknown": "Ocurrió un error inesperado",
       "empty": "El nombre no puede estar vacío",
-      "invalid_characters": "El nombre solo puede contener letras, números, espacios, guiones bajos y guiones",
       "name_too_long": "El nombre debe tener 50 caracteres o menos"
     },
     "abort": {
@@ -74,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Actualizar configuración de la instancia",
-        "description": "Modifica la configuración para esta instancia de asistente de IA.",
+        "title": "Seleccionar proveedor",
+        "description": "Elige el proveedor de IA para esta instancia. La integración se recargará después de guardar los cambios.",
         "data": {
+          "api_provider": "Proveedor de API"
+        }
+      },
+      "settings": {
+        "title": "Configuración de conexión y modelo",
+        "description": "Configura las credenciales de API y los parámetros del modelo. Los cambios tendrán efecto después de recargar la integración.",
+        "data": {
+          "api_key": "Clave API",
+          "api_endpoint": "URL del endpoint de API",
           "model": "Modelo de IA",
           "temperature": "Creatividad de la respuesta (0-2)",
           "max_tokens": "Longitud máxima de la respuesta (1-100000)",
           "request_interval": "Intervalo mínimo de solicitud (0.1-60 segundos)",
           "api_timeout": "Tiempo de espera de solicitud API en segundos (5-600)",
           "context_messages": "Número de mensajes anteriores a incluir en el contexto (1-20)",
-          "max_history_size": "Tamaño máximo del historial de conversación (1-100)"
+          "max_history_size": "Tamaño máximo del historial de conversación (1-100)",
+          "allow_local_network": "Permitir endpoints de red local (para proxies autoalojados)",
+          "disable_thinking": "Desactivar modo thinking/reasoning (Qwen /no_think, elimina bloques think, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -130,6 +136,18 @@
         "max_tokens": {
           "name": "Máx. Tokens",
           "description": "Longitud máxima de la respuesta (1-100000 tokens)"
+        },
+        "structured_output": {
+          "name": "Salida Estructurada",
+          "description": "Habilitar modo de salida JSON estructurada. Cuando está habilitado, la IA responderá con JSON válido que coincida con el esquema proporcionado."
+        },
+        "json_schema": {
+          "name": "Esquema JSON",
+          "description": "Esquema JSON que define la estructura de la respuesta esperada. Requerido cuando structured_output está habilitado."
+        },
+        "disable_thinking": {
+          "name": "Desactivar Thinking",
+          "description": "Desactivar el modo thinking/reasoning para esta solicitud. Anula la configuración de la integración."
         }
       }
     },
@@ -200,8 +218,7 @@
           "rate_limited": "Limitado por tasa",
           "maintenance": "Mantenimiento",
           "initializing": "Inicializando",
-          "retrying": "Reintentando",
-          "queued": "En cola"
+          "retrying": "Reintentando"
         },
         "state_attributes": {
           "question": {
@@ -293,6 +310,21 @@
           },
           "min_latency": {
             "name": "Latencia Mínima"
+          },
+          "last_model": {
+            "name": "Último modelo utilizado"
+          },
+          "last_timestamp": {
+            "name": "Hora de la última respuesta"
+          },
+          "instance_name": {
+            "name": "Nombre de instancia"
+          },
+          "normalized_name": {
+            "name": "Nombre normalizado"
+          },
+          "conversation_history": {
+            "name": "Historial de conversación"
           }
         }
       }

custom_components/ha_text_ai/translations/hi.json

@@ -14,7 +14,9 @@
           "request_interval": "अनुरोधों के बीच न्यूनतम समय (0.1-60 सेकंड)",
           "api_timeout": "एपीआई अनुरोध टाइमआउट सेकंड में (5-600)",
           "context_messages": "रखने के लिए संदर्भ संदेशों की संख्या (1-20)",
-          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)"
+          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)",
+          "allow_local_network": "स्थानीय नेटवर्क एंडपॉइंट की अनुमति दें (सेल्फ-होस्टेड प्रॉक्सी के लिए)",
+          "disable_thinking": "thinking/reasoning मोड बंद करें (Qwen /no_think, think ब्लॉक हटाता है, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -31,7 +33,9 @@
           "request_interval": "अनुरोधों के बीच न्यूनतम समय (0.1-60 सेकंड)",
           "api_timeout": "एपीआई अनुरोध टाइमआउट सेकंड में (5-600)",
           "context_messages": "रखने के लिए संदर्भ संदेशों की संख्या (1-20)",
-          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)"
+          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)",
+          "allow_local_network": "स्थानीय नेटवर्क एंडपॉइंट की अनुमति दें (सेल्फ-होस्टेड प्रॉक्सी के लिए)",
+          "disable_thinking": "thinking/reasoning मोड बंद करें (Qwen /no_think, think ब्लॉक हटाता है, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -42,6 +46,7 @@
       "name_exists": "इस नाम के साथ एक उदाहरण पहले से मौजूद है",
       "invalid_name": "अमान्य उदाहरण नाम",
       "invalid_auth": "प्रमाणीकरण विफल - अपनी एपीआई कुंजी की जांच करें",
+      "api_key_required": "प्रदाता या endpoint बदलते समय API कुंजी आवश्यक है",
       "invalid_api_key": "अमान्य एपीआई कुंजी - कृपया अपनी क्रेडेंशियल्स की पुष्टि करें",
       "cannot_connect": "एपीआई सेवा से कनेक्ट करने में विफल",
       "invalid_model": "चुना हुआ मॉडल उपलब्ध नहीं है",
@@ -55,7 +60,6 @@
       "invalid_instance": "अमान्य उदाहरण निर्दिष्ट किया गया",
       "unknown": "अप्रत्याशित त्रुटि हुई",
       "empty": "नाम खाली नहीं हो सकता",
-      "invalid_characters": "नाम में केवल अक्षर, अंक, रिक्त स्थान, अंडरस्कोर और हाइफन हो सकते हैं",
       "name_too_long": "नाम 50 अक्षरों या उससे कम होना चाहिए"
     },
     "abort": {
@@ -65,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "उदाहरण सेटिंग्स अपडेट करें",
-        "description": "इस एआई सहायक उदाहरण के लिए सेटिंग्स संशोधित करें।",
+        "title": "प्रदाता चुनें",
+        "description": "इस उदाहरण के लिए एआई प्रदाता चुनें। परिवर्तन सहेजने के बाद एकीकरण पुनः लोड होगा।",
         "data": {
+          "api_provider": "एपीआई प्रदाता"
+        }
+      },
+      "settings": {
+        "title": "कनेक्शन और मॉडल सेटिंग्स",
+        "description": "एपीआई क्रेडेंशियल और मॉडल पैरामीटर कॉन्फ़िगर करें। एकीकरण पुनः लोड होने के बाद परिवर्तन प्रभावी होंगे।",
+        "data": {
+          "api_key": "एपीआई कुंजी",
+          "api_endpoint": "एपीआई एंडपॉइंट यूआरएल",
           "model": "एआई मॉडल",
           "temperature": "प्रतिक्रिया की रचनात्मकता (0-2)",
           "max_tokens": "प्रतिक्रिया की अधिकतम लंबाई (1-100000)",
           "request_interval": "न्यूनतम अनुरोध अंतराल (0.1-60 सेकंड)",
           "api_timeout": "एपीआई अनुरोध टाइमआउट सेकंड में (5-600)",
           "context_messages": "संदर्भ में शामिल करने के लिए पिछले संदेशों की संख्या (1-20)",
-          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)"
+          "max_history_size": "अधिकतम बातचीत इतिहास आकार (1-100)",
+          "allow_local_network": "स्थानीय नेटवर्क एंडपॉइंट की अनुमति दें (सेल्फ-होस्टेड प्रॉक्सी के लिए)",
+          "disable_thinking": "thinking/reasoning मोड बंद करें (Qwen /no_think, think ब्लॉक हटाता है, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -121,6 +136,18 @@
         "max_tokens": {
           "name": "अधिकतम टोकन",
           "description": "प्रतिक्रिया की अधिकतम लंबाई (1-100000 टोकन)"
+        },
+        "structured_output": {
+          "name": "संरचित आउटपुट",
+          "description": "JSON संरचित आउटपुट मोड सक्षम करें। सक्षम होने पर, AI प्रदान किए गए स्कीमा से मेल खाने वाले वैध JSON के साथ प्रतिक्रिया देगा।"
+        },
+        "json_schema": {
+          "name": "JSON स्कीमा",
+          "description": "अपेक्षित प्रतिक्रिया की संरचना को परिभाषित करने वाला JSON स्कीमा। structured_output सक्षम होने पर आवश्यक।"
+        },
+        "disable_thinking": {
+          "name": "Thinking बंद करें",
+          "description": "इस अनुरोध के लिए thinking/reasoning मोड बंद करें। एकीकरण सेटिंग को ओवरराइड करता है।"
         }
       }
     },
@@ -191,8 +218,7 @@
           "rate_limited": "रेट सीमित",
           "maintenance": "रखरखाव",
           "initializing": "प्रारंभिककरण",
-          "retrying": "पुनः प्रयास कर रहा है",
-          "queued": "क्यू में"
+          "retrying": "पुनः प्रयास कर रहा है"
         },
         "state_attributes": {
           "question": {
@@ -284,9 +310,24 @@
           },
           "min_latency": {
             "name": "न्यूनतम विलंबता"
+          },
+          "last_model": {
+            "name": "अंतिम उपयोग किया गया मॉडल"
+          },
+          "last_timestamp": {
+            "name": "अंतिम प्रतिक्रिया समय"
+          },
+          "instance_name": {
+            "name": "इंस्टेंस नाम"
+          },
+          "normalized_name": {
+            "name": "सामान्यीकृत नाम"
+          },
+          "conversation_history": {
+            "name": "वार्तालाप इतिहास"
           }
         }
       }
     }
   }
-}
+}

custom_components/ha_text_ai/translations/it.json

@@ -1,15 +1,6 @@
 {
   "config": {
     "step": {
-      "provider": {
-        "title": "Seleziona fornitore AI",
-        "description": "Scegli quale fornitore di servizi AI utilizzare per questa istanza.",
-        "data": {
-          "api_provider": "Fornitore API",
-          "context_messages": "Numero di messaggi di contesto da mantenere (1-20)",
-          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)"
-        }
-      },
       "provider": {
         "title": "Impostazioni fornitore",
         "description": "Fornisci i dettagli di connessione per il tuo fornitore di AI scelto.",
@@ -23,7 +14,9 @@
           "request_interval": "Tempo minimo tra le richieste (0.1-60 secondi)",
           "api_timeout": "Timeout della richiesta API in secondi (5-600)",
           "context_messages": "Numero di messaggi di contesto da mantenere (1-20)",
-          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)"
+          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)",
+          "allow_local_network": "Consenti endpoint di rete locale (per proxy self-hosted)",
+          "disable_thinking": "Disabilita la modalità thinking/reasoning (Qwen /no_think, rimuove i blocchi think, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -40,7 +33,9 @@
           "request_interval": "Tempo minimo tra le richieste (0.1-60 secondi)",
           "api_timeout": "Timeout della richiesta API in secondi (5-600)",
           "context_messages": "Numero di messaggi di contesto da mantenere (1-20)",
-          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)"
+          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)",
+          "allow_local_network": "Consenti endpoint di rete locale (per proxy self-hosted)",
+          "disable_thinking": "Disabilita la modalità thinking/reasoning (Qwen /no_think, rimuove i blocchi think, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -51,6 +46,7 @@
       "name_exists": "Esiste già un'istanza con questo nome",
       "invalid_name": "Nome dell'istanza non valido",
       "invalid_auth": "Autenticazione fallita - controlla la tua chiave API",
+      "api_key_required": "La chiave API è obbligatoria quando si cambia provider o endpoint",
       "invalid_api_key": "Chiave API non valida - verifica le tue credenziali",
       "cannot_connect": "Impossibile connettersi al servizio API",
       "invalid_model": "Il modello selezionato non è disponibile",
@@ -64,7 +60,6 @@
       "invalid_instance": "Istanze specificata non valida",
       "unknown": "Si è verificato un errore imprevisto",
       "empty": "Il nome non può essere vuoto",
-      "invalid_characters": "Il nome può contenere solo lettere, numeri, spazi, trattini bassi e trattini",
       "name_too_long": "Il nome deve essere lungo 50 caratteri o meno"
     },
     "abort": {
@@ -74,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Aggiorna impostazioni dell'istanza",
-        "description": "Modifica le impostazioni per questa istanza di assistente AI.",
+        "title": "Seleziona fornitore",
+        "description": "Scegli il fornitore AI per questa istanza. L'integrazione verrà ricaricata dopo aver salvato le modifiche.",
         "data": {
+          "api_provider": "Fornitore API"
+        }
+      },
+      "settings": {
+        "title": "Impostazioni di connessione e modello",
+        "description": "Configura le credenziali API e i parametri del modello. Le modifiche avranno effetto dopo il ricaricamento dell'integrazione.",
+        "data": {
+          "api_key": "Chiave API",
+          "api_endpoint": "URL dell'endpoint API",
           "model": "Modello AI",
           "temperature": "Creatività della risposta (0-2)",
           "max_tokens": "Lunghezza massima della risposta (1-100000)",
           "request_interval": "Intervallo minimo di richiesta (0.1-60 secondi)",
           "api_timeout": "Timeout della richiesta API in secondi (5-600)",
           "context_messages": "Numero di messaggi precedenti da includere nel contesto (1-20)",
-          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)"
+          "max_history_size": "Dimensione massima della cronologia delle conversazioni (1-100)",
+          "allow_local_network": "Consenti endpoint di rete locale (per proxy self-hosted)",
+          "disable_thinking": "Disabilita la modalità thinking/reasoning (Qwen /no_think, rimuove i blocchi think, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -130,6 +136,18 @@
         "max_tokens": {
           "name": "Token massimi",
           "description": "Lunghezza massima della risposta (1-100000 token)"
+        },
+        "structured_output": {
+          "name": "Output Strutturato",
+          "description": "Abilita la modalità di output JSON strutturato. Quando abilitato, l'IA risponderà con JSON valido corrispondente allo schema fornito."
+        },
+        "json_schema": {
+          "name": "Schema JSON",
+          "description": "Schema JSON che definisce la struttura della risposta attesa. Richiesto quando structured_output è abilitato."
+        },
+        "disable_thinking": {
+          "name": "Disabilita Thinking",
+          "description": "Disabilita la modalità thinking/reasoning per questa richiesta. Sovrascrive l'impostazione dell'integrazione."
         }
       }
     },
@@ -200,8 +218,7 @@
           "rate_limited": "Limite di frequenza",
           "maintenance": "Manutenzione",
           "initializing": "Inizializzazione",
-          "retrying": "Riprova",
-          "queued": "In coda"
+          "retrying": "Riprova"
         },
         "state_attributes": {
           "question": {
@@ -293,6 +310,21 @@
           },
           "min_latency": {
             "name": "Latenza minima"
+          },
+          "last_model": {
+            "name": "Ultimo modello utilizzato"
+          },
+          "last_timestamp": {
+            "name": "Ora dell'ultima risposta"
+          },
+          "instance_name": {
+            "name": "Nome istanza"
+          },
+          "normalized_name": {
+            "name": "Nome normalizzato"
+          },
+          "conversation_history": {
+            "name": "Cronologia conversazione"
           }
         }
       }

custom_components/ha_text_ai/translations/ru.json

@@ -1,15 +1,6 @@
 {
   "config": {
     "step": {
-      "provider": {
-        "title": "Выбор провайдера ИИ",
-        "description": "Выберите сервис искусственного интеллекта для этого экземпляра.",
-        "data": {
-          "api_provider": "Провайдер API",
-          "context_messages": "Количество сохраняемых контекстных сообщений (1-20)",
-          "max_history_size": "Максимальный размер истории разговора (1-100)"
-        }
-      },
       "provider": {
         "title": "Настройки провайдера",
         "description": "Укажите параметры подключения для выбранного провайдера ИИ.",
@@ -23,7 +14,9 @@
           "request_interval": "Минимальный интервал между запросами (0.1-60 секунд)",
           "api_timeout": "Таймаут API-запроса в секундах (5-600)",
           "context_messages": "Количество сохраняемых контекстных сообщений (1-20)",
-          "max_history_size": "Максимальный размер истории разговора (1-100)"
+          "max_history_size": "Максимальный размер истории разговора (1-100)",
+          "allow_local_network": "Разрешить локальные сетевые адреса (для self-hosted прокси)",
+          "disable_thinking": "Отключить режим thinking/reasoning (Qwen /no_think, срезание блоков think, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -40,7 +33,9 @@
           "request_interval": "Минимальный интервал между запросами (0.1-60 секунд)",
           "api_timeout": "Таймаут API-запроса в секундах (5-600)",
           "context_messages": "Количество сохраняемых контекстных сообщений (1-20)",
-          "max_history_size": "Максимальный размер истории разговора (1-100)"
+          "max_history_size": "Максимальный размер истории разговора (1-100)",
+          "allow_local_network": "Разрешить локальные сетевые адреса (для self-hosted прокси)",
+          "disable_thinking": "Отключить режим thinking/reasoning (Qwen /no_think, срезание блоков think, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -51,6 +46,7 @@
       "name_exists": "Экземпляр с таким именем уже существует",
       "invalid_name": "Недопустимое имя экземпляра",
       "invalid_auth": "Ошибка аутентификации - проверьте API-ключ",
+      "api_key_required": "Необходимо ввести API-ключ при смене провайдера или эндпоинта",
       "invalid_api_key": "Недопустимый API-ключ - пожалуйста, проверьте учетные данные",
       "cannot_connect": "Не удалось подключиться к сервису API",
       "invalid_model": "Выбранная модель недоступна",
@@ -64,7 +60,6 @@
       "invalid_instance": "Указан некорректный экземпляр",
       "unknown": "Произошла непредвиденная ошибка",
       "empty": "Имя не может быть пустым",
-      "invalid_characters": "Имя может содержать только буквы, цифры, пробелы, подчеркивания и дефисы",
       "name_too_long": "Имя должно быть не длиннее 50 символов"
     },
     "abort": {
@@ -74,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Обновление настроек экземпляра",
-        "description": "Измените настройки для этого экземпляра ИИ-помощника.",
+        "title": "Выбор провайдера",
+        "description": "Выберите провайдера ИИ для этого экземпляра. Интеграция перезагрузится после сохранения изменений.",
         "data": {
+          "api_provider": "Провайдер API"
+        }
+      },
+      "settings": {
+        "title": "Настройки подключения и модели",
+        "description": "Настройте учётные данные API и параметры модели. Изменения вступят в силу после перезагрузки интеграции.",
+        "data": {
+          "api_key": "API-ключ",
+          "api_endpoint": "URL конечной точки API",
           "model": "Модель ИИ",
           "temperature": "Креативность ответа (0-2)",
           "max_tokens": "Максимальная длина ответа (1-100000)",
           "request_interval": "Минимальный интервал между запросами (0.1-60 секунд)",
           "api_timeout": "Таймаут API-запроса в секундах (5-600)",
           "context_messages": "Количество предыдущих сообщений для включения в контекст (1-20)",
-          "max_history_size": "Максимальный размер истории разговора (1-100)"
+          "max_history_size": "Максимальный размер истории разговора (1-100)",
+          "allow_local_network": "Разрешить локальные сетевые адреса (для self-hosted прокси)",
+          "disable_thinking": "Отключить режим thinking/reasoning (Qwen /no_think, срезание блоков think, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -130,6 +136,18 @@
         "max_tokens": {
           "name": "Максимум токенов",
           "description": "Максимальная длина ответа (1-100000 токенов)"
+        },
+        "structured_output": {
+          "name": "Структурированный вывод",
+          "description": "Включить режим структурированного JSON-вывода. При включении ИИ будет отвечать валидным JSON, соответствующим указанной схеме."
+        },
+        "json_schema": {
+          "name": "JSON Schema",
+          "description": "JSON-схема, определяющая структуру ожидаемого ответа. Обязательна при включении structured_output."
+        },
+        "disable_thinking": {
+          "name": "Отключить thinking",
+          "description": "Отключить режим thinking/reasoning для этого запроса. Переопределяет настройку интеграции."
         }
       }
     },
@@ -200,8 +218,7 @@
           "rate_limited": "Лимит запросов",
           "maintenance": "Техническое обслуживание",
           "initializing": "Инициализация",
-          "retrying": "Повторная попытка",
-          "queued": "В очереди"
+          "retrying": "Повторная попытка"
         },
         "state_attributes": {
           "question": {
@@ -293,6 +310,24 @@
           },
           "min_latency": {
             "name": "Минимальная задержка"
+          },
+          "last_model": {
+            "name": "Последняя использованная модель"
+          },
+          "last_timestamp": {
+            "name": "Время последнего ответа"
+          },
+          "instance_name": {
+            "name": "Имя экземпляра"
+          },
+          "normalized_name": {
+            "name": "Нормализованное имя"
+          },
+          "last_error": {
+            "name": "Последняя ошибка"
+          },
+          "conversation_history": {
+            "name": "История разговоров"
           }
         }
       }

custom_components/ha_text_ai/translations/sr.json

@@ -14,7 +14,9 @@
           "request_interval": "Минимално време између захтева (0.1-60 секунди)",
           "api_timeout": "Временско ограничење API захтева у секундама (5-600)",
           "context_messages": "Број контекстуалних порука које треба задржати (1-20)",
-          "max_history_size": "Максимална величина историје разговора (1-100)"
+          "max_history_size": "Максимална величина историје разговора (1-100)",
+          "allow_local_network": "Дозволи локалне мрежне адресе (за self-hosted проксије)",
+          "disable_thinking": "Онемогући thinking/reasoning режим (Qwen /no_think, уклања think блокове, Gemini 2.5 thinking_budget=0)"
         }
       },
       "user": {
@@ -31,7 +33,9 @@
           "request_interval": "Минимално време између захтева (0.1-60 секунди)",
           "api_timeout": "Временско ограничење API захтева у секундама (5-600)",
           "context_messages": "Број контекстуалних порука које треба задржати (1-20)",
-          "max_history_size": "Максимална величина историје разговора (1-100)"
+          "max_history_size": "Максимална величина историје разговора (1-100)",
+          "allow_local_network": "Дозволи локалне мрежне адресе (за self-hosted проксије)",
+          "disable_thinking": "Онемогући thinking/reasoning режим (Qwen /no_think, уклања think блокове, Gemini 2.5 thinking_budget=0)"
         }
       }
     },
@@ -42,6 +46,7 @@
       "name_exists": "Инстанца са овим именом већ постоји",
       "invalid_name": "Неважеће име инстанце",
       "invalid_auth": "Аутентификација није успела - проверите ваш API кључ",
+      "api_key_required": "API кључ је обавезан при промени провајдера или endpoint-а",
       "invalid_api_key": "Неважећи API кључ - молимо проверите ваше акредитиве",
       "cannot_connect": "Неуспело повезивање са API сервисом",
       "invalid_model": "Изабрани модел није доступан",
@@ -55,7 +60,6 @@
       "invalid_instance": "Неважећа инстанца је назначена",
       "unknown": "Дошло је до неочекиване грешке",
       "empty": "Име не може бити празно",
-      "invalid_characters": "Име може садржавати само слова, цифре, размаке, подцрта и цртице",
       "name_too_long": "Име мора бити 50 знакова или мање"
     },
     "abort": {
@@ -65,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "Ажурирајте подешавања инстанце",
-        "description": "Измените подешавања за ову AI асистент инстанцу.",
+        "title": "Изаберите провајдера",
+        "description": "Изаберите AI провајдера за ову инстанцу. Интеграција ће се поново учитати након чувања измена.",
         "data": {
+          "api_provider": "API провајдер"
+        }
+      },
+      "settings": {
+        "title": "Подешавања везе и модела",
+        "description": "Конфигуришите API акредитиве и параметре модела. Промене ће ступити на снагу након поновног учитавања интеграције.",
+        "data": {
+          "api_key": "API кључ",
+          "api_endpoint": "URL API крајње тачке",
           "model": "AI модел",
           "temperature": "Креативност одговора (0-2)",
           "max_tokens": "Максимална дужина одговора (1-100000)",
           "request_interval": "Минимално време између захтева (0.1-60 секунди)",
           "api_timeout": "Временско ограничење API захтева у секундама (5-600)",
           "context_messages": "Број претходних порука које треба укључити у контекст (1-20)",
-          "max_history_size": "Максимална величина историје разговора (1-100)"
+          "max_history_size": "Максимална величина историје разговора (1-100)",
+          "allow_local_network": "Дозволи локалне мрежне адресе (за self-hosted проксије)",
+          "disable_thinking": "Онемогући thinking/reasoning режим (Qwen /no_think, уклања think блокове, Gemini 2.5 thinking_budget=0)"
         }
       }
     }
@@ -121,6 +136,18 @@
         "max_tokens": {
           "name": "Максимални токени",
           "description": "Максимална дужина одговора (1-100000 токена)"
+        },
+        "structured_output": {
+          "name": "Структурисани излаз",
+          "description": "Омогући JSON структурисани излаз. Када је омогућено, AI ће одговарати валидним JSON-ом који одговара датој шеми."
+        },
+        "json_schema": {
+          "name": "JSON шема",
+          "description": "JSON шема која дефинише структуру очекиваног одговора. Обавезна када је structured_output омогућен."
+        },
+        "disable_thinking": {
+          "name": "Онемогући thinking",
+          "description": "Онемогући thinking/reasoning режим за овај захтев. Надмашује поставку интеграције."
         }
       }
     },
@@ -187,12 +214,11 @@
           "ready": "Спремно",
           "processing": "Обрада",
           "error": "Грешка",
-          "disconnected": "Прекључено",
+          "disconnected": "Искључено",
           "rate_limited": "Ограничење захтева",
           "maintenance": "Одржавање",
           "initializing": "Инициализује се",
-          "retrying": "Покушава поново",
-          "queued": "У реду"
+          "retrying": "Покушава поново"
         },
         "state_attributes": {
           "question": {
@@ -284,9 +310,24 @@
           },
           "min_latency": {
             "name": "Минимална латенција"
+          },
+          "last_model": {
+            "name": "Последњи коришћени модел"
+          },
+          "last_timestamp": {
+            "name": "Време последњег одговора"
+          },
+          "instance_name": {
+            "name": "Назив инстанце"
+          },
+          "normalized_name": {
+            "name": "Нормализовани назив"
+          },
+          "conversation_history": {
+            "name": "Историја разговора"
           }
         }
       }
     }
   }
-}
+}

custom_components/ha_text_ai/translations/zh.json

@@ -14,7 +14,9 @@
           "request_interval": "请求之间的最小时间（0.1-60秒）",
           "api_timeout": "API请求超时时间（5-600秒）",
           "context_messages": "保留的上下文消息数量（1-20）",
-          "max_history_size": "最大对话历史大小（1-100）"
+          "max_history_size": "最大对话历史大小（1-100）",
+          "allow_local_network": "允许本地网络端点（用于自托管代理）",
+          "disable_thinking": "禁用思考/推理模式（Qwen /no_think，移除 think 块，Gemini 2.5 thinking_budget=0）"
         }
       },
       "user": {
@@ -31,7 +33,9 @@
           "request_interval": "请求之间的最小时间（0.1-60秒）",
           "api_timeout": "API请求超时时间（5-600秒）",
           "context_messages": "保留的上下文消息数量（1-20）",
-          "max_history_size": "最大对话历史大小（1-100）"
+          "max_history_size": "最大对话历史大小（1-100）",
+          "allow_local_network": "允许本地网络端点（用于自托管代理）",
+          "disable_thinking": "禁用思考/推理模式（Qwen /no_think，移除 think 块，Gemini 2.5 thinking_budget=0）"
         }
       }
     },
@@ -42,6 +46,7 @@
       "name_exists": "具有此名称的实例已存在",
       "invalid_name": "无效的实例名称",
       "invalid_auth": "身份验证失败 - 检查您的API密钥",
+      "api_key_required": "更改提供商或端点时需要输入 API 密钥",
       "invalid_api_key": "无效的API密钥 - 请验证您的凭据",
       "cannot_connect": "无法连接到API服务",
       "invalid_model": "所选模型不可用",
@@ -55,7 +60,6 @@
       "invalid_instance": "指定的实例无效",
       "unknown": "发生意外错误",
       "empty": "名称不能为空",
-      "invalid_characters": "名称只能包含字母、数字、空格、下划线和连字符",
       "name_too_long": "名称必须少于50个字符"
     },
     "abort": {
@@ -65,16 +69,27 @@
   "options": {
     "step": {
       "init": {
-        "title": "更新实例设置",
-        "description": "修改此AI助手实例的设置。",
+        "title": "选择提供者",
+        "description": "选择此实例的AI提供者。保存更改后集成将重新加载。",
         "data": {
+          "api_provider": "API提供者"
+        }
+      },
+      "settings": {
+        "title": "连接和模型设置",
+        "description": "配置API凭据和模型参数。更改将在集成重新加载后生效。",
+        "data": {
+          "api_key": "API密钥",
+          "api_endpoint": "API端点URL",
           "model": "AI模型",
           "temperature": "响应创造力（0-2）",
           "max_tokens": "最大响应长度（1-100000）",
           "request_interval": "最小请求间隔（0.1-60秒）",
           "api_timeout": "API请求超时时间（5-600秒）",
           "context_messages": "要包含在上下文中的先前消息数量（1-20）",
-          "max_history_size": "最大对话历史大小（1-100）"
+          "max_history_size": "最大对话历史大小（1-100）",
+          "allow_local_network": "允许本地网络端点（用于自托管代理）",
+          "disable_thinking": "禁用思考/推理模式（Qwen /no_think，移除 think 块，Gemini 2.5 thinking_budget=0）"
         }
       }
     }
@@ -121,6 +136,18 @@
         "max_tokens": {
           "name": "最大标记数",
           "description": "响应的最大长度（1-100000个标记）"
+        },
+        "structured_output": {
+          "name": "结构化输出",
+          "description": "启用JSON结构化输出模式。启用后，AI将以符合提供的模式的有效JSON进行响应。"
+        },
+        "json_schema": {
+          "name": "JSON模式",
+          "description": "定义预期响应结构的JSON模式。启用structured_output时必需。"
+        },
+        "disable_thinking": {
+          "name": "禁用思考",
+          "description": "为此请求禁用思考/推理模式。覆盖集成级别的设置。"
         }
       }
     },
@@ -191,8 +218,7 @@
           "rate_limited": "速率限制",
           "maintenance": "维护中",
           "initializing": "初始化中",
-          "retrying": "重试中",
-          "queued": "排队中"
+          "retrying": "重试中"
         },
         "state_attributes": {
           "question": {
@@ -284,9 +310,24 @@
           },
           "min_latency": {
             "name": "最小延迟"
+          },
+          "last_model": {
+            "name": "最近使用的模型"
+          },
+          "last_timestamp": {
+            "name": "最近响应时间"
+          },
+          "instance_name": {
+            "name": "实例名称"
+          },
+          "normalized_name": {
+            "name": "规范化名称"
+          },
+          "conversation_history": {
+            "name": "对话历史"
           }
         }
       }
     }
   }
-}
+}

custom_components/ha_text_ai/utils.py

@@ -0,0 +1,258 @@
+"""
+Utility functions for HA Text AI integration.
+
+@license: MIT (https://opensource.org/licenses/MIT)
+@author: SMKRV
+@github: https://github.com/smkrv/ha-text-ai
+@source: https://github.com/smkrv/ha-text-ai
+"""
+from __future__ import annotations
+
+import hashlib
+import ipaddress
+import logging
+import socket
+from typing import Any
+from urllib.parse import urlparse
+
+import aiohttp
+from aiohttp.abc import AbstractResolver
+
+from homeassistant.const import CONF_API_KEY
+from homeassistant.core import HomeAssistant
+
+_LOGGER = logging.getLogger(__name__)
+
+def normalize_name(name: str) -> str:
+    """Normalize name to conform to HA naming convention using underscores.
+
+    If the input collapses to an empty string (all non-alphanumeric or
+    all underscores), fall back to a short hash of the original so that
+    downstream entity IDs never end with a trailing underscore.
+    """
+    normalized = ''.join(c if c.isalnum() or c == '_' else '_' for c in name)
+    normalized = '_'.join(filter(None, normalized.split('_'))).lower()
+    if not normalized:
+        digest = hashlib.sha256(name.encode("utf-8", errors="replace")).hexdigest()[:8]
+        normalized = f"instance_{digest}"
+    return normalized
+
+def safe_log_data(
+    data: dict[str, Any],
+    sensitive_keys: tuple[str, ...] = (CONF_API_KEY,),
+) -> dict[str, Any]:
+    """Filter sensitive keys from data for safe logging."""
+    return {k: "***" if k in sensitive_keys else v for k, v in data.items()}
+
+class _RestrictedIPError(ValueError):
+    """Raised when an IP address is in a restricted range."""
+
+def _check_ip_restricted(addr: ipaddress.IPv4Address | ipaddress.IPv6Address) -> bool:
+    """Check if an IP address is in a restricted range."""
+    return (
+        addr.is_private
+        or addr.is_reserved
+        or addr.is_loopback
+        or addr.is_link_local
+        or addr.is_multicast
+        or addr.is_unspecified
+    )
+
+def _is_cloud_metadata_or_unsafe(
+    addr: ipaddress.IPv4Address | ipaddress.IPv6Address,
+) -> bool:
+    """Block link-local and cloud instance-metadata addresses.
+
+    These must be blocked even in allow_local_network mode:
+    - IPv4 link-local (169.254.0.0/16) covers AWS/GCP/Azure IMDS 169.254.169.254.
+    - IPv6 link-local (fe80::/10).
+    - Multicast and unspecified addresses.
+    Without this check, a self-hosted HA running on a cloud VM could be
+    tricked into exfiltrating cloud credentials via IMDS.
+    """
+    return addr.is_multicast or addr.is_unspecified or addr.is_link_local
+
+class _PinnedResolver(AbstractResolver):
+    """aiohttp resolver that returns pre-validated IPs for a single hostname.
+
+    Why: prevents DNS-rebinding attacks. After validate_endpoint has
+    confirmed the hostname resolves to a safe IP, we pin that IP in the
+    resolver used by the aiohttp session. aiohttp then skips its own
+    DNS lookup on each request and uses the pinned IP, closing the
+    TOCTOU gap between validation and actual HTTP call.
+    """
+
+    def __init__(
+        self,
+        pinned: dict[str, list[tuple[str, int]]],
+    ) -> None:
+        self._pinned = pinned
+
+    async def resolve(
+        self,
+        host: str,
+        port: int = 0,
+        family: int = socket.AF_INET,
+    ) -> list[dict[str, Any]]:
+        entries = self._pinned.get(host.lower())
+        if entries is None:
+            # Every request on a pinned session must target the validated
+            # host. Resolving anything else means a request escaped the pin
+            # (a new call site or a config bug) — fail closed rather than
+            # fall back to live, unvalidated DNS.
+            raise OSError(f"Refusing to resolve unpinned host: {host}")
+        return [
+            {
+                "hostname": host,
+                "host": ip,
+                "port": port or default_port,
+                "family": _family_for(ip),
+                "proto": 0,
+                "flags": 0,
+            }
+            for ip, default_port in entries
+        ]
+
+    async def close(self) -> None:
+        """Nothing to close; the resolver holds only the pinned map."""
+
+def _family_for(ip: str) -> int:
+    """Return AF_INET or AF_INET6 based on the IP literal."""
+    try:
+        return socket.AF_INET6 if ":" in ip else socket.AF_INET
+    except Exception:
+        return socket.AF_INET
+
+async def resolve_hostname_ips(
+    hass: HomeAssistant,
+    hostname: str,
+) -> list[str]:
+    """Resolve hostname to all its IPs via the event-loop-safe executor.
+
+    Returns a list of IP strings (may contain both IPv4 and IPv6).
+    Raises ValueError on resolution failure.
+    """
+    try:
+        addrinfos = await hass.async_add_executor_job(
+            socket.getaddrinfo, hostname, None
+        )
+    except socket.gaierror as err:
+        raise ValueError(f"Cannot resolve hostname: {hostname}") from err
+    ips = []
+    seen: set[str] = set()
+    for _family, _type, _proto, _canonname, sockaddr in addrinfos:
+        ip = sockaddr[0]
+        if ip not in seen:
+            seen.add(ip)
+            ips.append(ip)
+    if not ips:
+        raise ValueError(f"No IPs for hostname: {hostname}")
+    return ips
+
+def create_pinned_session(
+    endpoint: str,
+    resolved_ips: list[str],
+) -> aiohttp.ClientSession:
+    """Create an isolated aiohttp session with pinned DNS and no cookie jar.
+
+    Addresses two issues at once:
+    - DNS rebinding: aiohttp will reuse the pinned IPs from validate_endpoint
+      rather than re-resolving the hostname on each request.
+    - Cookie pollution: DummyCookieJar prevents cookies from leaking between
+      this integration and other HA components sharing the same domain.
+
+    Built directly on aiohttp: HA's async_create_clientsession always
+    injects its own pooled connector and rejects a caller-supplied one,
+    so a custom resolver cannot go through the helper. The caller owns
+    the session and must close it. Requests must pass
+    allow_redirects=False so a 3xx response cannot route past the pinned
+    resolver to an unvalidated host.
+    """
+    parsed = urlparse(endpoint)
+    hostname = (parsed.hostname or "").lower()
+    port = parsed.port or (443 if parsed.scheme == "https" else 80)
+    pinned: dict[str, list[tuple[str, int]]] = {
+        hostname: [(ip, port) for ip in resolved_ips]
+    }
+    connector = aiohttp.TCPConnector(resolver=_PinnedResolver(pinned))
+    return aiohttp.ClientSession(
+        connector=connector,
+        cookie_jar=aiohttp.DummyCookieJar(),
+    )
+
+async def validate_endpoint(
+    hass: HomeAssistant,
+    endpoint: str,
+    *,
+    allow_local: bool = False,
+) -> tuple[str, list[str]]:
+    """Validate API endpoint URL for security and pin resolved IPs.
+
+    Ensures HTTPS-only and blocks private/reserved IP ranges (SSRF protection).
+    When allow_local is True, permits private IPs and HTTP scheme for self-hosted proxies.
+    Uses async DNS resolution to avoid blocking the event loop.
+
+    Returns: (validated_endpoint_without_trailing_slash, resolved_ips).
+    The resolved IPs are intended for pinning in aiohttp resolver via
+    create_pinned_session(), closing the DNS-rebinding TOCTOU gap.
+
+    Raises:
+        ValueError: If the endpoint fails validation.
+    """
+    parsed = urlparse(endpoint)
+
+    if allow_local:
+        if parsed.scheme not in ("https", "http"):
+            raise ValueError("Only HTTPS and HTTP endpoints are allowed")
+    else:
+        if parsed.scheme not in ("https",):
+            raise ValueError("Only HTTPS endpoints are allowed")
+
+    hostname = parsed.hostname
+    if not hostname:
+        raise ValueError("Invalid endpoint URL: no hostname")
+
+    resolved_ips: list[str] = []
+
+    # Collect and check all resolved IPs (or IP literal directly).
+    def _collect(ips: list[str]) -> None:
+        for ip in ips:
+            if ip not in resolved_ips:
+                resolved_ips.append(ip)
+
+    try:
+        addr = ipaddress.ip_address(hostname)
+        _is_ip_literal = True
+    except ValueError:
+        addr = None
+        _is_ip_literal = False
+
+    if _is_ip_literal:
+        _collect([hostname])
+    else:
+        try:
+            addrinfos = await hass.async_add_executor_job(
+                socket.getaddrinfo, hostname, None
+            )
+        except socket.gaierror as err:
+            raise ValueError(f"Cannot resolve hostname: {hostname}") from err
+        _collect([sockaddr[0] for (*_, sockaddr) in addrinfos])
+
+    if not resolved_ips:
+        raise ValueError(f"No IPs resolved for hostname: {hostname}")
+
+    # Validate each resolved IP against the selected policy.
+    for ip_str in resolved_ips:
+        ip_obj = ipaddress.ip_address(ip_str)
+        if allow_local:
+            if _is_cloud_metadata_or_unsafe(ip_obj):
+                raise ValueError(
+                    "Link-local/metadata/multicast addresses are not allowed"
+                )
+        else:
+            if _check_ip_restricted(ip_obj):
+                raise _RestrictedIPError(
+                    "Private/reserved IP addresses are not allowed"
+                )
+
+    return endpoint.rstrip("/"), resolved_ips

structure.md

@@ -1,22 +1,27 @@
 ```
-ha_text_ai/
+custom_components/ha_text_ai/
 ├── __init__.py
 ├── api_client.py
 ├── config_flow.py
 ├── const.py
 ├── coordinator.py
-├── icons
-│   ├── dark_icon.png
-│   ├── dark_icon@2x.png
-│   ├── dark_logo.png
-│   ├── dark_logo@2x.png
-│   ├── icon.png
-│   ├── icon@2x.png
-│   ├── logo.png
-│   └── logo@2x.png
-├── manifest.json
+├── history.py
+├── metrics.py
+├── providers.py
 ├── sensor.py
 ├── services.yaml
+├── strings.json
+├── utils.py
+├── icons
+│   ├── dark_icon.png
+│   ├── dark_icon@2x.png
+│   ├── dark_logo.png
+│   ├── dark_logo@2x.png
+│   ├── icon.png
+│   ├── icon@2x.png
+│   ├── logo.png
+│   └── logo@2x.png
+├── manifest.json
 └── translations
     ├── de.json
     ├── en.json