WorkBuddy Connector MCP server for self-hosted ASR transcription (stdio, auth_mode: token)
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"workbuddy-mcp-asr","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# ASR Transcribe Connector(asr-transcribe) WorkBuddy 连接器:把本机音频文件上传到用户**自托管**的 ASR 转写服务,返回转写文字。 按 06 Connector 开发规范(MCP 方案)+ 06D 附录C(MCP 用户自填 Token 模式)实现: - MCP Server:Python + FastMCP(stdio 本地子进程),依赖 `mcp[cli]` + `httpx` - 鉴权:`auth_mode: "token"`,凭证通过 token-schema 表单由用户填写,仅存储在本机, 连接时以 `${VAR}` 占位符注入 stdio 子进程的 env ## 文件结构 ``` asr-transcribe/ ├── connector-meta.json # 元信息(source: asr-transcribe, type: mcp, auth_mode: token) ├── mcp.json # MCP 连接配置(stdio + env 注入 ASR_URL / ASR_USER / ASR_PASS) ├── token-schema.json # 用户自填凭证表单定义 ├── skills/ │ └── SKILL.md # 教 AI 何时/如何调用转写工具 ├── server/ │ └── asr_mcp_server.py # MCP Server 本体 ├── icon.svg # 图标(麦克风) └── README.md ``` ## 对接的后端 API MCP Server 只对接自托管 ASR 服务的两个端点: | 端点 | 方法 | 说明 | |---|---|---| | `/` | POST (multipart, 字段 `file`) | 上传音频,触发转写,结果落盘到服务端监听目录 | | `/api/jobs` | GET | JSON 列表,每项含 `name` / `status` / `text` | ## 提供的 MCP 工具 - `transcribe_audio(file_path)` — 上传音频并轮询 `/api/jobs`(间隔 3s、最多 60 次),返回转写文本 - `list_recent_jobs(limit=20)` — 返回最近的转写记录列表 ## 凭证(token-schema.json) | 字段 | 类型 | 必填 | 说明 | |---|---|---|---| | ASR_URL | text | ✅ | 服务基础地址,默认 `http://127.0.0.1:8899`(参考实现的默认端口) | | ASR_USER | text | ✅ | Basic Auth 用户名 | | ASR_PASS | password | ✅ | Basic Auth 密码 | 三项凭证仅存储在本机(`~/.workbuddy/connectors/asr-transcribe/` 下),由服务提供方发放账号。 ## 本地测试步骤 1. **裸测后端 API**(本机真服务在 `127.0.0.1:8899`,无 basic auth,可直接验证对接逻辑): ```bash curl -s http://127.0.0.1:8899/api/jobs | python3 -m json.tool ``` 2. **准备 venv**(已就绪可跳过): ```bash python3 -m venv .venv .venv/bin/pip install "mcp[cli]" httpx ``` 3. **无令牌/缺凭证时的友好错误**: ```bash env -i PATH="$PATH" .venv/bin/python -c " import asyncio, server.asr_mcp_server as s print(asyncio.run(s.list_recent_jobs.fn(limit=5)) if hasattr(s.list_recent_jobs,'fn') else 'see test')" ``` (详见 `server/test_smoke.py`,缺 env 应返回「缺少连接配置:ASR_URL、ASR_USER、ASR_PASS…」) 4. **带凭证实测转写**(用 8899 本地服务): ```bash ASR_URL=ht…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.