yt-dj-mcp
Syncs YouTube Music liked songs, analyzes them for DJ metadata like BPM and key, and enables creating playlists from previews.
README
yt-dj-mcp
Local YouTube Music DJ crate MCP server.
The server syncs liked songs from YouTube Music, stores approximate DJ metadata such as BPM, key, Camelot key, and genre, then creates playlists only from saved previews.
Who This Is For
Use this if you have a large YouTube Music liked-song library and want a local, agent-friendly way to turn it into DJ crates. The workflow is deliberately preview-first: inspect candidates locally, then create playlists only after you approve a saved preview.
What It Can Do
- Sync YouTube Music liked songs into local SQLite.
- Analyze short local audio sections with
yt-dlp,ffmpeg, and Essentia. - Store BPM, musical key, Camelot key, confidence, and source.
- Search tracks by BPM/key/genre.
- Save playlist previews before writing to YouTube Music.
- Create YouTube Music playlists from reviewed previews.
Safety
- Uses OAuth by default.
- Stores secrets in
~/.config/yt-dj-mcp/. - Never prints credential file contents.
- Creates playlists only from saved previews.
- Browser/HAR auth is an explicit fallback and stores session headers in
browser.json; keep it local and never commit it. - Runtime state is local-only. Do not publish
library.sqlite, HAR files, browser headers, OAuth tokens, or downloaded audio.
Quick Start
git clone https://github.com/YOUR_USERNAME/yt-dj-mcp.git
cd yt-dj-mcp
python -m venv .venv
. .venv/bin/activate
pip install -e ".[dev,audio]"
pytest -q
Add the MCP server to your client:
[mcp_servers.yt_dj_mcp]
command = "/absolute/path/to/yt-dj-mcp/.venv/bin/yt-dj-mcp"
args = ["serve"]
OAuth Setup
Create a Google Cloud OAuth client that can be used by ytmusicapi, then run:
yt-dj-mcp setup-auth
yt-dj-mcp auth-status
If OAuth token creation succeeds but YouTube Music requests return HTTP 400, use the explicit local browser fallback:
ytmusicapi browser --file ~/.config/yt-dj-mcp/browser.json
chmod 600 ~/.config/yt-dj-mcp/browser.json
yt-dj-mcp auth-status
If copying request headers from DevTools is unreliable, export a Chrome Network HAR and import it:
yt-dj-mcp import-har /path/to/music.youtube.com.har
yt-dj-mcp auth-status
browser.json contains browser session headers. Keep it local and never commit it.
Commands
yt-dj-mcp setup-auth
yt-dj-mcp auth-status
yt-dj-mcp import-har /path/to/music.youtube.com.har
yt-dj-mcp enrich-tracks --limit 25
yt-dj-mcp reset-auth
yt-dj-mcp serve
MCP tools exposed by the server:
auth_statussetup_auth_instructionssync_liked_songslist_liked_songsenrich_tracksfind_dj_candidatescreate_playlist_previewget_playlist_previewcreate_youtube_music_playlistupdate_track_metadata
Audio Metadata Enrichment
enrich-tracks fills missing BPM/key/Camelot metadata by downloading a short
audio section with yt-dlp, analyzing it with Essentia, then caching the
analysis under ~/.config/yt-dj-mcp/audio-metadata-cache/.
Required local commands/packages:
ffmpeg -version
python -m yt_dlp --version
python -c "import essentia; print(essentia.__version__)"
The command is best run in small batches:
yt-dj-mcp enrich-tracks --limit 10
yt-dj-mcp enrich-tracks --limit 50
If yt-dlp or Essentia is missing, enrichment will leave tracks in the
missing count instead of crashing the server. Use this only for music you are
allowed to process locally; downloading streamed audio may be restricted by the
service's terms.
First Crate Workflow
- Run
yt-dj-mcp setup-auth. - Run
yt-dj-mcp auth-status. - Add the MCP server to Codex or Claude.
- Call
sync_liked_songs. - Run
yt-dj-mcp enrich-tracks --limit 25or call MCPenrich_tracks. - Call
find_dj_candidates. - Call
create_playlist_preview. - Review with
get_playlist_preview. - Create the YouTube Music playlist with
create_youtube_music_playlist.
Suggested Agent Prompt
After installing the MCP server, try:
Sync my YouTube Music liked songs, enrich 50 tracks with BPM/key metadata, then
create 12-track DJ playlist previews grouped by compatible BPM and Camelot key.
Review previews before asking the agent to create real YouTube Music playlists.
Limitations
- BPM/key values are estimates. Use your ears before playing out.
- YouTube Music auth can be brittle. OAuth is preferred, while browser/HAR auth is a local fallback when OAuth does not work.
- Audio analysis depends on
ffmpeg,yt-dlp, and Essentia availability. - This project does not manage playback and does not provide destructive library operations.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.