Skip to main content
The ČSFD API includes a built-in Model Context Protocol (MCP) server, allowing you to use ČSFD data directly within Large Language Models like Claude Desktop.

What is MCP?

Model Context Protocol (MCP) is a standard protocol that allows AI assistants to access external data sources and tools. With MCP, Claude Desktop and other AI assistants can:
  • Search for movies and creators on ČSFD
  • Retrieve detailed information about films
  • Access user ratings and reviews
  • Get cinema screenings
All in real-time, directly from natural language conversations.

Features

The ČSFD MCP server provides the following capabilities:

Search

Search for movies, TV series, creators, and users

Movie Details

Get comprehensive movie information including ratings, cast, and plot

Creator Info

Access biographies and filmographies

User Data

Retrieve user ratings and reviews

Quick Start with Claude Desktop

The easiest way to use the MCP server is with Claude Desktop.
1

Locate your Claude Desktop config

Find the configuration file:
~/Library/Application Support/Claude/claude_desktop_config.json
2

Add ČSFD MCP server configuration

Add the following to your claude_desktop_config.json:
{
  "mcpServers": {
    "csfd": {
      "command": "npx",
      "args": ["-y", "node-csfd-api", "mcp"]
    }
  }
}
If you already have other MCP servers configured, just add the csfd entry to your existing mcpServers object.
3

Restart Claude Desktop

Close and reopen Claude Desktop for the changes to take effect.
4

Test it out

Try asking Claude:
  • “Search for Inception on ČSFD”
  • “Get details about Quentin Tarantino”
  • “What are the top rated movies by Christopher Nolan?”

Available MCP Tools

The ČSFD MCP server exposes 6 tools that AI assistants can use: Description: Search for movies, TV series, or people on ČSFD.cz Input:
  • query (string): Search query (movie title, actor name, etc.)
Returns: Lists of matching movies, TV series, creators, and users with their IDs
This is usually the first tool to use - it returns IDs that other tools need.

2. Get Movie

Description: Retrieve detailed information about a specific movie or series Input:
  • id (number): ČSFD movie ID (obtained from search)
Returns: Comprehensive movie data including:
  • Title, year, rating, and poster
  • Plot descriptions and trivia
  • Genres and origins
  • Cast and crew
  • VOD platforms
  • Premieres and related movies

3. Get Creator

Description: Get information about a creator (actor, director, etc.) Input:
  • id (number): ČSFD creator ID
Returns:
  • Name and biography
  • Birthday and birthplace
  • Photo
  • Complete filmography

4. Get User Ratings

Description: Retrieve movie ratings from a ČSFD user profile Input:
  • user (string | number): User ID or username
  • page (number, optional): Page number
  • allPages (boolean, optional): Fetch all pages
  • allPagesDelay (number, optional): Delay between requests (ms)
  • excludes (array, optional): Film types to exclude
  • includesOnly (array, optional): Film types to include
Returns: Array of rated movies with user ratings (0-5 stars)

5. Get User Reviews

Description: Retrieve detailed reviews written by a ČSFD user Input: Same as Get User Ratings Returns: Array of reviews with full text and ratings

6. Get Cinemas

Description: Get cinema screenings for a specific district Input:
  • district (string | number): District ID or name
  • period (enum): Time period (today, tomorrow, weekend, week, month)
Returns: List of cinemas with their screenings and showtimes

Usage Examples

Here are some example prompts you can use with Claude Desktop:

Built-in Prompts

The MCP server includes pre-built prompts for common tasks:

Actor Top Rated

Description: Finds and ranks the best movies of a specific actor or creator Usage:
Find top rated movies by Mads Mikkelsen
This prompt automatically:
  1. Searches for the person
  2. Retrieves their profile
  3. Identifies top 5 highest-rated movies
  4. Returns a formatted list

Running the MCP Server Manually

You can also run the MCP server directly from the command line:
npx node-csfd-api mcp

Configuration for Other AI Assistants

While Claude Desktop is the primary target, the MCP server can work with any MCP-compatible AI assistant.

Generic MCP Client Configuration

{
  "name": "csfd",
  "command": "npx",
  "args": ["-y", "node-csfd-api", "mcp"],
  "transport": "stdio"
}

Troubleshooting

Solution:
  1. Verify your claude_desktop_config.json is correctly formatted
  2. Make sure you restarted Claude Desktop after making changes
  3. Check that npx is available in your PATH:
    npx --version
    
Solution:
  1. Check your internet connection
  2. Verify the ČSFD website is accessible
  3. Try running the MCP server manually to see error messages:
    npx node-csfd-api mcp
    
Solution:
  • The MCP server scrapes data from ČSFD.cz in real-time, so some delay is normal
  • For faster responses, consider running your own REST API server with caching
Solution:
  • When fetching all pages of user ratings/reviews, use the allPagesDelay parameter
  • Avoid making too many requests in a short time
  • Consider implementing your own caching layer

Advanced Configuration

Using a Local Build

For development or if you want to modify the MCP server:
claude_desktop_config.json
{
  "mcpServers": {
    "csfd-local": {
      "command": "node",
      "args": ["/path/to/node-csfd-api/dist/bin/mcp-server.mjs"]
    }
  }
}

Environment Variables

You can pass environment variables to the MCP server:
{
  "mcpServers": {
    "csfd": {
      "command": "npx",
      "args": ["-y", "node-csfd-api", "mcp"],
      "env": {
        "NODE_ENV": "production"
      }
    }
  }
}

Privacy & Security

The MCP server does not collect, store, or transmit any user data. All requests go directly from your machine to ČSFD.cz.
  • No data is logged or stored
  • No intermediary servers are involved
  • All scraping happens in real-time from your local machine
  • Respects ČSFD’s terms of service

Next Steps

TypeScript Types

Explore the type definitions for MCP responses

REST API

Run your own ČSFD API server

Build docs developers (and LLMs) love