MCP Server

A lightweight Model Context Protocol (MCP) server for managing movie/TV series watchlists using TMDB data. Built with FastAPI, raw SQL, and Pydantic.

Features

• Search, discover, and fetch movie/TV metadata via TMDB API
• Track watchlist items with statuses: PlanToWatch, Watching, Watched, Dropped, OnHold
• Store film info, metadata, and genres in SQLite using raw SQL
• JSON-RPC-style MCP interface over HTTP for structured command handling

Tech Stack

• Python, FastAPI, Pydantic
• SQLite
• TMDB API

Sample

Run Locally

API Server

git clone https://github.com/yourusername/mcp-server.git
cd mcp-server
pip install -r requirements.txt

export TMDB_API_KEY=your_tmdb_key_here
uvicorn main:app --reload

NOTE: Create '.env' and add TMDB_API_KEY variable there

MCP Config file (Claude Config compatible)

• UV (Recommended) Note: Install and configure (uv)[https://docs.astral.sh/uv/guides/tools/] for mcp before excuting

{
  "mcpServers": {
    "Movie MCP": {
      "command": "uv",
      "args": [
        "run",
        "mcp", "run", "[path_to_mcp_tmdb]/mcp_server.py"
      ]
    }
  }
}

• Hacky way

{
  "mcpServers": {
    
    "Movie MCP": {
      "command": "[path_to_mcp_tmdb]/setup.sh",
      "args": []
    }
  }
}

Use any Claude mcp config client and add above config. Use Console-chat-gpt CLI based chat MCP Client

Alternatively use (mcptools)[https://github.com/f/mcptools/]

• List tools

cd [path_to_movie_mcp]
mcptools tools uv run mcp run mcp_server.py

• Search Movie

cd [path_to_movie_mcp]
mcptools call search_tmdb --params '{"query":"final destination", "type":"movie"}' uv run mcp run mcp_server.py