mcp-music-forge
MCP server for downloading and processing audio from SoundCloud and other sources, with tools for URL probing, download enqueueing, and job status tracking.
README
MCP Music Forge
English | Русский
A scalable MCP plugin/service for downloading and processing audio: SoundCloud (starting point), then YouTube/Yandex Music/Spotify; transcoding, tagging and cover art embedding, HTTP API, MCP tools, audio workers.
Quickstart
Variant A: Docker Compose (recommended)
cp .env.example .env
make upb # build and start the stack
Endpoints:
- API: http://localhost:8033
- Admin: http://localhost:8033/admin
- MCP HTTP: http://localhost:8033/mcp
Management:
make logs # logs
make ps # container status
make up # just up
make down # stop and remove
Variant B: Local (uv)
make install # create .venv, install deps, copy .env
source .venv/bin/activate
make lint # ruff + black + mypy
make test # mypy + pytest
# run API
uvicorn api.main:app --reload
# MCP (stdio)
python -m mcp_music_forge.mcp_app
API Examples
# health check
curl -s http://localhost:8033/health | jq
# {"status": "ok"}
# enqueue download (SoundCloud URL with allowed download per ToU)
curl -s -X POST 'http://localhost:8033/download?url=https://soundcloud.com/artist/track' | jq
# {"job_id": "abc123", "status": "queued"}
# check job status
curl -s http://localhost:8033/jobs/<job_id> | jq
MCP Tools
probe_url: provider detection and downloadability check.enqueue_download: create/duplicate a job, put it in the queue.get_job_status: status, artifacts, file links as MCP resources.- Resources:
forge://jobs/<job_id>/{original|final}/<filename>(file bytes).
Project Overview
- MCP Server (
mcp_music_forge/): job management, resource provider, and MCP tools. - HTTP API (
api/):POST /download,GET /jobs/{id},/health, admin interface. - Providers (
providers/): adapters for sources (starting with SoundCloud). - Transcoder (
transcoder/): a wrapper aroundffmpeg. - Storage (
storage/): local FS (can be replaced with S3, etc.). - Queue (
core/services/queue.py): ARQ + Redis; a wrapper inworkers/.
Documentation
Legal Notes
- SoundCloud provider respects ToU: we only download if the track is downloadable (
downloadable/download_url). - Cookie file support is available, but use it strictly within the service's rules.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。