From 1ad8aeeaa711058d7658297a001a759172105943 Mon Sep 17 00:00:00 2001
From: hrz <1710360675@qq.com>
Date: Thu, 22 May 2025 11:22:16 +0800
Subject: [PATCH] =?UTF-8?q?add=EF=BC=9ACustomTTS=E6=8F=90=E4=BE=9B?=
 =?UTF-8?q?=E4=B8=80=E4=B8=AA=E7=A4=BA=E4=BE=8B?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 main/xiaozhi-server/config.yaml | 25 ++++++++++++++++---------
 1 file changed, 16 insertions(+), 9 deletions(-)

diff --git a/main/xiaozhi-server/config.yaml b/main/xiaozhi-server/config.yaml
index 05756cd7..44845df8 100644
--- a/main/xiaozhi-server/config.yaml
+++ b/main/xiaozhi-server/config.yaml
@@ -675,17 +675,24 @@ TTS:
     speed: 1
     output_dir: tmp/
   CustomTTS:
-    # 自定义的TTS接口服务，请求参数可自定义
-    # 要求接口使用GET方式请求，并返回音频文件
+    # 自定义的TTS接口服务，请求参数可自定义，可接入众多TTS服务
+    # 以本地部署的KokoroTTS为例
+    # 如果只有cpu运行：docker run -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-cpu:latest
+    # 如果只有gpu运行：docker run --gpus all -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-gpu:latest
+    # 要求接口使用POST方式请求，并返回音频文件
     type: custom
-    url: "http://127.0.0.1:9880/tts"
+    method: POST
+    url: "http://127.0.0.1:8880/v1/audio/speech"
     params: # 自定义请求参数
-      # text: "{prompt_text}" # {prompt_text}会被替换为实际的提示词内容
-      # speaker: jok老师
-      # speed: 1
-      # foo: bar
-      # testabc: 123456
+      input: "{prompt_text}"
+      response_format: "mp3"
+      download_format: "mp3"
+      voice: "zf_xiaoxiao"
+      lang_code: "z"
+      return_download_link: true
+      speed: 1
+      stream: false
     headers: # 自定义请求头
       # Authorization: Bearer xxxx
-    format: wav # 接口返回的音频格式
+    format: mp3 # 接口返回的音频格式
     output_dir: tmp/
\ No newline at end of file