What2Watch

What2Watch MCP enables your AI to access real-time entertainment data from live TV schedules, free streaming channels, and major SVOD services, delivering truly personalized recommendations and discovery across platforms. It connects your assistant to up-to-date availability and watching options so you can ask for things like what to watch tonight, what's on a specific channel, or to discover hidden gems based on mood and time.

How to use

You use What2Watch MCP by connecting your AI assistant to the MCP endpoint, then issuing discovery or search commands. The MCP exposes several tools that let you find content, check schedules, view details, and see what’s trending. You will receive structured results including titles, genres, ratings, brief plots, and where to watch.

Typical usage patterns include: get personalized recommendations with mood and watch time, search for titles or actors, retrieve full content details, check live schedules, and see what’s trending across platforms. Use the example prompts and queries to guide natural language requests to your AI assistant.

How to install

Prerequisites: you need a runtime environment for your MCP client, typically Node.js or Python, and a client that can communicate with MCP servers via the Model Context Protocol (SSE transport). Ensure you have network access to reach the MCP endpoint.

Step 1: Identify the MCP endpoint you will connect to. In this case, you will use the live MCP endpoint URL for What2Watch.

Step 2: Configure your MCP client to connect to the endpoint. The server uses SSE for transport and requires standard OAuth 2.1 with PKCE for authentication when applicable.

Step 3: Begin making requests to the available tools, such as get_recommendation, search_content, get_schedule, get_content_details, get_trending, and get_available_sources, to start discovering content and schedules.

Configuration and usage notes

Endpoint: https://what2watch.live/sse. Transport: SSE. Authentication: OAuth 2.1 with PKCE. The MCP exposes a set of tools that return structured results including content items with availability data across platforms.

If you integrate with multiple assistants (Claude, ChatGPT, etc.), follow their specific steps to add a custom or app-based connector pointing to the same endpoint. Use the exact URL and ensure your app has the required permissions to access the endpoint.

Available tools and how they work

What2Watch MCP provides six primary tools for discovery and content access. Each tool returns a structured response with content items, availability, and guidance for follow-up actions.

interface GetRecommendationParams {
  location?: string
  languages?: string[]
  gender?: string
  age_group?: string
  platforms?: string[]
  channels?: string[]
  liked_media?: string[]
  disliked_media?: string[]
  watched_media?: string[]
  media_type?: string
  genre?: string
  mood?: string
  watch_time?: string
  surprise_me?: boolean
  limit?: number
}

Example prompts

Discovery prompts include what_to_watch, surprise_me, recommend_by_mood, and similar_to. Search prompts include find_title, find_by_actor, find_by_genre, recent_releases. Trending prompts cover whats_trending, trending_on_platform, and top_rated. TV schedule prompts include whats_on_tv and whats_on_channel. Occasion-based prompts help tailor recommendations for movie nights, family viewing, or date night.

Available tools

get_recommendation

Fetch personalized movie and TV show suggestions based on user preferences and mood.

search_content

Search for titles by keywords, genres, year, location, language, platforms, and viewing status.

get_content_details

Retrieve full details for a specific content item, including plot, cast, and streaming availability.

get_schedule

Obtain live or upcoming TV schedules with filters for time range, channels, and genres.

get_trending

See what content is currently popular across platforms and genres.

get_available_sources

List all supported streaming services and channels with availability information.