ani-mcp

A smart MCP server for AniList that understands your anime and manga taste - not just raw API calls.

What makes this different

Most AniList integrations mirror the API 1:1. ani-mcp adds an intelligence layer on top:

• Taste profiling - builds a model of your preferences from your completed list
• Personalized picks - "what should I watch next?" based on your taste, mood, and seasonal lineup
• Compatibility - compare taste between two users
• Sequel alerts - find sequels airing this season for shows you've finished
• Watch order - franchise chain traversal for long-running series
• Session planning - "I have 90 minutes, what should I watch?" from your current list
• Year in review - your watching/reading stats wrapped up

Plus the essentials: search, details, trending, seasonal browsing, list management, social features, and community recommendations. All search and browse tools support pagination for browsing beyond the first page of results.

Try it in 30 seconds

No account needed. Works with any MCP-compatible client.

Claude Desktop

Add to your config file (Settings > Developer > Edit Config or ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "anilist": {
      "command": "npx",
      "args": ["-y", "ani-mcp"]
    }
  }
}

Restart Claude Desktop after saving.

Alternatively, download ani-mcp.mcpb from the latest release and install via Settings > Extensions.

Claude Code

claude mcp add ani-mcp -- npx -y ani-mcp

Personalized features

Add your username for recommendations, taste profiling, and list management:

{
  "mcpServers": {
    "anilist": {
      "command": "npx",
      "args": ["-y", "ani-mcp"],
      "env": {
        "ANILIST_USERNAME": "your_username"
      }
    }
  }
}

For write operations (updating progress, scoring, list edits), also add ANILIST_TOKEN. See Environment Variables for details.

Environment Variables

Tools

Search & Discovery

Lists & Stats

Intelligence

Cards

Info

Social

Analytics

Write (requires ANILIST_TOKEN)

Resources

MCP resources provide context to your AI assistant without needing a tool call. Clients that support resources can automatically include this information in conversations.

Prompts

Pre-built conversation starters that clients can offer as quick actions.

Examples

Here are some things you can ask your AI assistant once ani-mcp is connected:

"What should I watch next?" Uses anilist_pick to analyze your completed list, build a taste profile, and recommend titles from your Planning list ranked by how well they match your preferences.

"I want something dark and psychological" Uses anilist_pick with mood filtering to boost titles matching that vibe and penalize mismatches.

"What's good this season?" Uses anilist_pick with source: SEASONAL to rank currently airing anime against your taste profile.

"I have 90 minutes, what should I watch tonight?" Uses anilist_session to pick from your currently watching list and fill a time budget with the best-matching episodes.

"Any sequels airing for stuff I've finished?" Uses anilist_sequels to cross-reference your completed list with this season's lineup.

"What order do I watch Fate in?" Uses anilist_watch_order to traverse the franchise relation graph and produce a numbered viewing order.

"Compare my taste with username123" Uses anilist_compare to find shared titles, compute a compatibility score, and highlight biggest disagreements.

"Why would I like Vinland Saga?" Uses anilist_explain to score a title against your taste profile, breaking down genre affinity and theme alignment.

"Show me my anime year in review" Uses anilist_wrapped to summarize everything you watched in a given year.

Mood config

anilist_pick and anilist_session accept a freeform mood string. Built-in keywords include: dark, chill, hype, action, romantic, funny, brainy, sad, scary, epic, wholesome, intense, mystery, fantasy, scifi, trippy, nostalgic, artistic, competitive, cozy.

To add or override keywords, set ANILIST_MOOD_CONFIG as a JSON object:

{
  "ANILIST_MOOD_CONFIG": "{\"cozy\":{\"boost\":[\"Slice of Life\",\"Iyashikei\"],\"penalize\":[\"Horror\"]},\"mykeyword\":{\"boost\":[\"Romance\"],\"penalize\":[]}}"
}

Each key is a mood keyword mapping to { boost: string[], penalize: string[] } arrays of AniList genres and tags.

Privacy

See PRIVACY.md for details. In short: ani-mcp runs locally, sends requests only to the AniList API, stores nothing, and collects no analytics.

Docker

docker build -t ani-mcp .
docker run -e ANILIST_USERNAME=your_username ani-mcp

Runs on port 3000 with HTTP Stream transport by default.

Build from Source

git clone https://github.com/gavxm/ani-mcp.git
cd ani-mcp
npm install
npm run build
npm test

Support

Bug reports and feature requests: GitHub Issues

License

MIT