@cyanheads/tmdb-mcp-server

Overview

Movie and TV metadata powered by The Movie Database (TMDB). Search titles and people by name, pull full credits and filmographies, discover titles by filtered criteria, track trending content, and check region-aware streaming availability 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 — the three resources are convenience wrappers over the detail service methods, so tool-only clients lose nothing.

Related MCP server: Movie MCP Server

Capability reference

tmdb_search_titles tool

• multi mode (default) mixes movies, shows, and people, each result tagged with media_type; movie/tv/person restrict to one type and enable type-specific ranking

• Optional year filter (movie/tv modes), language override, include_adult toggle, and page for paging past the first 20

• Each result carries id, media_type, title/name, release_year, overview, vote_average, resolved genre_names, and the relevant poster/profile URL

• An empty result set is a valid answer, returned with recovery guidance — not an error

tmdb_get_movie tool

• Folds credits, videos, recommendations, similar, keywords, external ids, and release dates into one call via append_to_response — trim the append array to shrink the payload (e.g. ["credits"] for cast only)

• US theatrical certification extracted from the release_dates namespace; top-billed cast and key crew (Director/Writer/Screenplay/Producer); YouTube trailers with watch URLs

• Does not include streaming availability — that is region-specific; use tmdb_get_watch_providers

tmdb_get_show tool

• Same append_to_response set as tmdb_get_movie, with content_ratings (US TV rating) in place of release_dates

• Adds season summaries, creators, networks, and the last/next episode to air

• Pass a season_number from seasons[] to tmdb_get_season for the episode list

tmdb_get_season tool

• Per-episode names, air dates, runtimes, vote averages, still URLs, and guest stars (embedded per episode)

• Plus the season's regular recurring cast (distinct from per-episode guest stars)

• series_id is echoed from the input — the TMDB season endpoint does not return it. Season 0 is "Specials"

tmdb_get_person tool

• Biography, birth/death dates, place of birth, known-for department, aliases, gender label

• combined_credits split into cast_credits and crew_credits, recency-ordered (most recent first) and capped to 50 per list — the pre-cap totals are reported and truncation is disclosed in the enrichment

• IMDb id plus the extended cross-platform id set (Wikidata, social handles) for chaining to other servers

tmdb_discover_titles tool

• Filter by with_genres/without_genres, exact year or a release_date_gte/lte window, vote_average range, vote_count_gte floor, with_cast/with_crew (movie), with_networks (tv), with_watch_providers + watch_region, with_original_language, and runtime range

• Sort by popularity, revenue, vote average, vote count, or release date. Pair vote_average.desc with vote_count_gte (~100–1000) so a 10.0-from-3-votes title does not dominate

• with_cast/with_crew are movie-only and with_networks is tv-only on TMDB; the tool accepts them for both media_type values and no-ops the inapplicable ones with a notice

• with_watch_providers requires a watch_region — streaming availability is region-specific. Omitting the region returns a typed region_required error

tmdb_get_trending tool

• media_type: all (default) mixes movies, shows, and people, each tagged with media_type; or restrict to one type

• time_window: day (more volatile) or week (default, steadier)

• Results are the same ranked summary cards as search and discover, up to 20 per page, with page for paging

• An empty result set returns with recovery guidance, not an error

tmdb_get_watch_providers tool

• Returns flatrate (subscription), rent, buy, ads (ad-supported free), and free provider lists with logo URLs, plus the TMDB JustWatch-backed link — the supported path to actual deep links

• A watch_region (ISO 3166-1 alpha-2) is required: availability is region-specific and there is no global answer; the response always carries a region caveat

• An empty result for a region is a valid "not available to stream here" answer, not an error

• Provider ids in the result feed back into tmdb_discover_titles with_watch_providers

tmdb://movie/{movie_id} resource

• Same enriched record as tmdb_get_movie, with the full append_to_response set — application/json

• movie_id comes from tmdb_search_titles or tmdb_discover_titles

• Typed movie_not_found error when TMDB returns 404 for the id

tmdb://tv/{series_id} resource

• Same enriched record as tmdb_get_show, with the full append_to_response set — application/json

• series_id comes from tmdb_search_titles or tmdb_discover_titles

• Typed show_not_found error when TMDB returns 404 for the id

tmdb://person/{person_id} resource

• Same enriched record as tmdb_get_person — application/json

• person_id comes from tmdb_search_titles (mode "person")

• Typed person_not_found error when TMDB returns 404 for the id

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.

TMDB-specific:

• Single typed client for the TMDB v3 REST API, authenticated with a v4 Read Access Token over Authorization: Bearer

• Image *_path fields resolved to full https://image.tmdb.org/t/p/{size}{path} URLs — null paths omit the field rather than emitting a broken URL

• genre_ids[] resolved to genre names from per-media_type maps cached at startup, so list results carry readable genres without an extra call

• Detail tools collapse credits, videos, recommendations, similar, keywords, and external ids into one HTTP call via append_to_response

• Region-aware streaming availability backed by JustWatch — region is always an explicit input, never assumed

Agent-friendly output:

• TMDB attribution on every response — surfaced in enrichment.attribution so it reaches both the structuredContent and content[] client surfaces

• Provenance and pagination on list responses — total-result counts, page/total-pages, and an effective-filter notice so agents can reason about what was actually queried

• Empty results and empty provider regions return normally with recovery guidance, not as errors — "no titles matched" and "not streamable here" are real answers an agent must act on

• Region and truncation caveats are explicit — watch-provider results always state the region; capped filmographies disclose the pre-cap totals

Getting started

A TMDB API Read Access Token is required. Create a free TMDB account, then copy the API Read Access Token (a v4 JWT) from your API settings — this server authenticates v3 endpoints with Authorization: Bearer <token>, not the legacy ?api_key= query parameter.

Add the following to your MCP client configuration file.

{
  "mcpServers": {
    "tmdb-mcp-server": {
      "type": "stdio",
      "command": "bunx",
      "args": ["@cyanheads/tmdb-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info",
        "TMDB_API_KEY": "your-tmdb-read-access-token"
      }
    }
  }
}

Or with npx (no Bun required):

{
  "mcpServers": {
    "tmdb-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@cyanheads/tmdb-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info",
        "TMDB_API_KEY": "your-tmdb-read-access-token"
      }
    }
  }
}

Or with Docker:

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

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

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

Refer to "your MCP client configuration file" generically — different clients use different config paths, and this server isn't client-specific.

Prerequisites

• Bun v1.4.0 or higher (or Node.js v24+).

• A TMDB API Read Access Token (v4 JWT) — free from your TMDB API settings.

Installation

1. Clone the repository:

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

2. Navigate into the directory:

cd tmdb-mcp-server

3. Install dependencies:

bun install

4. Configure environment:

cp .env.example .env
# edit .env and set TMDB_API_KEY

Configuration

All configuration is validated at startup via Zod schemas. Key environment variables:

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

Running the server

Local development

• Build and run:

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

• Run checks and tests:

  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 tmdb-mcp-server .
docker run --rm -e TMDB_API_KEY=your-token -p 3010:3010 tmdb-mcp-server

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

Project structure

Development guide

See AGENTS.md (or CLAUDE.md, the same content) 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 the TMDB API: validate raw → normalize to the domain type → return the output schema; never fabricate missing fields (null poster paths omit the URL, unknown genre ids are dropped)

Attribution

This product uses the TMDB API but is not endorsed or certified by TMDB. Streaming availability data is provided by JustWatch via TMDB and is region-specific.

Contributing

Issues are welcome. Run checks and tests before submitting:

bun run devcheck
bun run test

License

Apache-2.0 — see LICENSE for details.