MCPcopy Index your code
hub / github.com/aahl/qwen-asr2api

github.com/aahl/qwen-asr2api @main

Chat with this repo
repository ↗ · DeepWiki ↗ · + Follow
12 symbols 28 edges 2 files 0 documented · 0% updated 41d ago★ 75
What it actually does AI analysis from the code graph — generated when you open this
loading…
README

🎤 Qwen ASR

Install / 安装

🐳 Docker compose

mkdir /opt/asr2api
cd /opt/asr2api
wget https://raw.githubusercontent.com/aahl/qwen-asr2api/refs/heads/main/docker-compose.yml
docker compose up -d

🐳 Docker run

docker run -d \
  --name asr2api \
  --restart=unless-stopped \
  -p 8820:80 \
  ghcr.nju.edu.cn/aahl/qwen-asr2api:main

🏠 Home Assistant OS Add-on

  1. 添加加载项仓库
  2. 打开 HomeAssistant,点击左侧菜单的 配置 (Settings) -> 加载项 (Add-ons)
  3. 点击右下角的 加载项商店 (Add-on Store)
  4. 点击右上角的三个点 -> 仓库 (Repositories)
  5. 在输入框填入:https://gitee.com/hasscc/addons, 点击添加 添加加载项仓库

  6. 安装加载项

  7. 刷新页面,找到并点击 Qwen ASR
  8. 点击 安装 (Install)
  9. 启动并设置开机启动

💻 Usage / 使用

🌐 CURL调用示例

curl --request POST \
  --url http://localhost:8820/v1/audio/transcriptions \
  --form model=qwen3-asr \
  --form file='@audio.wav'

注意:如果你用 curl -F/--form,不要手动设置 Content-Type: multipart/form-data,让 curl 自己生成带 boundary 的请求头。否则服务端会因为缺少 boundary 而拒绝请求。

🏠 Home Assistant

  1. 安装 AI Conversation 集成

    点击这里 一键安装,安装完记得重启HA

  2. 添加 AI Conversation 服务,配置模型提供商

    服务商: 自定义; 接口: http://4e0de88e-qwen-asr/v1; 密钥留空

  3. 添加STT模型
  4. 配置语音助手

🤖 Models / 模型

  • qwen3-asr
  • qwen3-asr:itn 启用逆文本标准化

🔗 Links / 相关链接

  • 默认转发目标:https://qwen-qwen3-asr-demo.ms.show
  • 说明:本项目当前是把远端 Gradio ASR Demo 包装成 OpenAI 风格接口,不是本地离线推理。
  • https://linux.do/t/topic/1367480

Core symbols most depended-on inside this repo

api_post
called by 2
asr2api/__init__.py
main
called by 2
asr2api/__init__.py
upload_file
called by 1
asr2api/__init__.py
run_inference
called by 1
asr2api/__init__.py
check_auth
called by 1
asr2api/__init__.py
create_app
called by 1
asr2api/__init__.py
init_session
called by 0
asr2api/__init__.py
on_cleanup
called by 0
asr2api/__init__.py

Shape

Function 11
Class 1

Languages

Python100%

Modules by API surface

asr2api/__init__.py12 symbols

For agents

$ claude mcp add qwen-asr2api \
  -- python -m otcore.mcp_server <graph>

⬇ download graph artifact