Overview

Anime and manga data from AniList, Jikan (MyAnimeList), and Kitsu. Search titles, pull full detail with side-by-side AniList and MAL scores, walk a franchise's watch order, check the airing schedule, and look up characters, voice actors, and studio filmographies from any MCP client. Runs as a stdio process or a local Streamable HTTP server.

Tools

Resources

All resource data is also reachable via tools. Use anime_search_media to discover AniList IDs before fetching the resource URI.

Capability reference

anime_search_media _tool

• Free-text title search with AniList primary. When AniList has no match at all for a query-only search, it falls back to Jikan (MAL) and maps the rows to AniList IDs, honoring include_adult. The fallback runs only at per_page 25 or less (Jikan's page size); above that, or if Jikan is unavailable, the AniList empty page comes back with a notice instead of an error
• Filter by genre, tag, season/year, format (TV, MOVIE, OVA, MANGA, NOVEL, etc.), and status (RELEASING, FINISHED, etc.); up to 5 sort values
• Needs at least one criterion; blank or whitespace query/genre/tag values count as absent, and a sort other than SEARCH_MATCH alone browses the catalog (missing_criteria otherwise)
• Default sort: SEARCH_MATCH with a query, POPULARITY_DESC without one
• Adult content gated behind explicit include_adult: true (default off)
• Pagination via page and per_page (max 50), driven by has_next_page; total_results is exact and appears only on the final page. AniList serves the first 5,000 results — the last reachable page carries a notice saying so, and deeper pages fail with page_depth_exceeded

anime_get_media _tool

• AniList full-detail query supplemented in parallel by Jikan (MAL score) and Kitsu (streaming links with sub/dub language flags) via Promise.allSettled; data_sources flags which succeeded
• AniList and MAL scores surfaced side by side — never blended — plus AniList weighted average/popularity and MAL rank/popularity
• Tags carry an is_spoiler flag from AniList's isGeneralSpoiler; the formatted text view hides spoiler and adult tags while the full array stays in structured output
• Streaming links: Kitsu primary (with per-platform sub/dub language lists), AniList externalLinks fallback when Kitsu has none
• not_found when the AniList ID doesn't resolve — use anime_search_media first

anime_get_relations _tool

• Multi-hop BFS over AniList's relation graph, up to max_depth hops (default 2, max 4)
• Entries split into main (source, adaptation, prequel, sequel — canonical story) and supplementary (side story, spin-off, compilation, and similar) via watch_order_category
• Each entry includes season_year and episode/chapter count for context
• not_found when the root AniList ID doesn't resolve

anime_get_schedule _tool

• Two modes: season (all anime airing in a season/year — both season and season_year required) and upcoming (next episode per airing title within days_ahead, 7 when omitted, max 30)
• invalid_season when mode is season but season/season_year is missing; conflicting_inputs when a mode gets the other mode's fields (upcoming with season/season_year, season with days_ahead)
• Adult titles excluded by default (include_adult); pagination via page/per_page (max 50), driven by has_next_page. In season mode total_results is exact and appears only on the final page; AniList serves the first 5,000 entries — the last reachable page carries a notice saying so, and deeper pages fail with page_depth_exceeded
• Airing timestamps are UTC ISO 8601, with time_until_airing_seconds for countdowns

anime_find_characters _tool

• Three lookup modes: id (media → cast), character_name, or voice_actor_name — exactly one per call. None fails missing_identifier; more than one fails conflicting_inputs. Names are trimmed, and a blank name counts as absent
• By-media mode supports a language filter over AniList's StaffLanguage enum (JAPANESE, ENGLISH, KOREAN, etc.); language with a name search fails conflicting_inputs
• Cast list capped at per_page (max 25); a capped page returns truncated: true with next-page guidance
• media_not_found / not_found distinguish an invalid media ID from a name search with no match

anime_get_recommendations _tool

• Returns AniList recommendations, adding Jikan/MAL vote counts when the same recommendation appears in both sources; sources identifies each contribution
• Scores stay separate — anilist_rating and jikan_votes are never blended into one figure
• Optional liked_aspects free-text field is echoed back unmodified, for the caller to contextualize picks
• AniList page capped at per_page (max 25); a capped page returns truncated: true with next-page guidance
• not_found when the source AniList ID doesn't resolve

anime_get_rankings _tool

• Three modes: top (all-time by score), trending (current week), seasonal (current or specified season/year, sorted by popularity)
• Filterable by genre, tag (an AniList tag name such as Isekai), and format in every mode; blank values count as absent. Adult content excluded by default (include_adult)
• seasonal takes season and season_year together, or neither for the current season (invalid_season otherwise); in top/trending they restrict the ranking, and season_label names the applied filter
• Pagination via page/per_page (max 50), driven by has_next_page; each entry carries a 1-based rank. total_results is exact and appears only on the final page; AniList serves the first 5,000 entries — the last reachable page carries a notice saying so, and deeper pages fail with page_depth_exceeded

anime_get_studio _tool

• Look up by name (search, trimmed) or id (direct AniList studio ID) — exactly one: neither fails missing_identifier, both fail conflicting_inputs; a blank name counts as absent
• Filmography sortable by POPULARITY_DESC (default), SCORE_DESC, START_DATE_DESC, or START_DATE
• One row per distinct title on a page, with is_main_studio set when any of the studio's credits on it is a main-studio credit
• not_found when neither the name search nor the ID lookup resolves
• Pagination via page/per_page (max 25, AniList's page size for a studio's titles), driven by has_next_page; total_titles is exact only when the whole filmography fits on page 1

anime://media/{id} _resource

• Compact media record as application/json — title variants, synopsis, scores, genres, streaming count, cover image
• id comes from anime_search_media or anime_get_media
• A flatter subset of anime_get_media — no tags, studios, streaming links, or relations

Features

Built on @cyanheads/mcp-ts-core: stdio and Streamable HTTP transports, pluggable auth (none / jwt / oauth), swappable storage (in-memory, filesystem, Supabase, Cloudflare KV/R2/D1), structured logging with optional OpenTelemetry tracing.

Anime/manga-specific:

• Three-source architecture: AniList GraphQL (primary), Jikan v4 REST (MAL scores + recommendations), Kitsu JSON:API (streaming links with sub/dub language detail)
• ID reconciliation via AniList's idMal bridge — no cross-source ID guessing
• Franchise relation graph traversal with heuristic ordering by relation-type priority, then season year
• Rate-limit-aware service layer: AniList 30 req/30s with automatic backoff; Jikan 350ms floor between calls
• Keyless by design — all three sources are public and require no API credentials

Agent-friendly output:

• Dual scores surfaced separately — AniList meanScore (0–100) and MAL score (0–10) with population size (scored_by) so agents can reason about weight; never blended into a composite
• Spoiler safety — tags carry an is_spoiler flag from AniList's isGeneralSpoiler; the formatted text view hides spoiler and adult tags while the full array stays in structured output
• Supplement provenance — anime_get_media.data_sources reports whether AniList, MAL/Jikan, and Kitsu data was retrieved; an unavailable supplement leaves its flag false
• UTC timestamps throughout; season labels echoed (WINTER 2024) to avoid the winter/spring/summer/fall boundary footgun

Getting started

No API keys required — all three upstream sources (AniList, Jikan, Kitsu) are keyless public APIs.

Add the following to your MCP client configuration file:

{
  "mcpServers": {
    "anime-mcp-server": {
      "type": "stdio",
      "command": "bunx",
      "args": ["@cyanheads/anime-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info"
      }
    }
  }
}

Or with npx (no Bun required):

{
  "mcpServers": {
    "anime-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@cyanheads/anime-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info"
      }
    }
  }
}

Or with Docker:

{
  "mcpServers": {
    "anime-mcp-server": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "MCP_TRANSPORT_TYPE=stdio",
        "ghcr.io/cyanheads/anime-mcp-server:latest"
      ]
    }
  }
}

For Streamable HTTP, set the transport and start the server:

MCP_TRANSPORT_TYPE=http MCP_HTTP_PORT=3010 bun run start:http
# Server listens at http://localhost:3010/mcp

Prerequisites

• Bun v1.4.0 or higher (or Node.js v24+). No API keys needed.

Installation

1. Clone the repository:

git clone https://github.com/cyanheads/anime-mcp-server.git

2. Navigate into the directory:

cd anime-mcp-server

3. Install dependencies:

bun install

4. Configure environment (optional):

cp .env.example .env
# edit .env if you want to change transport or log level

Configuration

No server-specific env vars are required. All framework variables are optional with sensible defaults.

See .env.example for the full list of optional overrides.

Running the server

Local development

• Build and run:

  sh

  # One-time build
  bun run rebuild
  
  # Run the built server
  bun run start:stdio
  # or
  bun run start:http
  

• Run checks and tests:

  sh

  bun run devcheck   # Lint, format, typecheck, security
  bun run test       # Vitest test suite
  bun run lint:mcp   # Validate MCP definitions against spec
  

Docker

docker build -t anime-mcp-server .
docker run --rm -p 3010:3010 anime-mcp-server

The Dockerfile defaults to HTTP transport, stateless session mode, and logs to /var/log/anime-mcp-server. OpenTelemetry peer dependencies are installed by default — build with --build-arg OTEL_ENABLED=false to omit them.

Project structure

Development guide

See CLAUDE.md/AGENTS.md for development guidelines and architectural rules. The short version:

• Handlers throw, framework catches — no try/catch in tool logic
• Use ctx.log for request-scoped logging, ctx.state for tenant-scoped storage
• Register new tools and resources via the barrels in src/mcp-server/*/definitions/index.ts
• Wrap external API calls: validate raw → normalize to domain type → return output schema; never fabricate missing fields
• AniList is primary — supplement failures (Jikan, Kitsu) degrade gracefully via Promise.allSettled

Contributing

Issues are welcome. Run checks and tests before submitting:

bun run devcheck
bun run test

License

Apache-2.0 — see LICENSE for details.