YouTube Watch Later MCP Server

The YouTube Watch Later MCP Server bridges the gap between AI assistants and YouTube’s “Watch Later” functionality. By exposing a lightweight, authenticated tool that pulls video URLs from a user‑defined playlist over an adjustable time window, it lets developers add dynamic content discovery into conversational workflows. Rather than hard‑coding a list of videos or scraping the site, the server queries the YouTube Data API v3 securely via OAuth2 and returns only the URLs that match the requested timeframe. This eliminates repetitive manual effort and ensures that the assistant always references the most recent additions.

For developers building AI‑powered content curation or recommendation systems, this server is a turnkey solution. It eliminates the need to manage YouTube API credentials, handle pagination, or parse complex response objects. The tool accepts a single optional parameter, , making it trivial to ask the assistant for “what did I add yesterday?” or “give me last week’s watch‑later videos.” The output is a clean array of links, ready to be passed into downstream prompts or displayed in UI panels. Because the server handles authentication internally, developers can embed it directly into their MCP configuration files and rely on a single refresh token that never expires unless revoked.

Key capabilities include:

• Time‑window filtering – retrieve videos added within any configurable number of days, enabling trend analysis or daily summaries.
• OAuth2 security – secure access to private YouTube data without exposing API keys or user credentials in code.
• Simplicity – a single tool, straightforward JSON response, and no need for external scripting or manual API calls.
• Extensibility – built on the MCP SDK, it can be easily expanded to include additional metadata (titles, thumbnails) or combined with other MCP servers for richer workflows.

Typical use cases span from personal productivity tools that remind users of recently saved videos, to content teams that aggregate fresh material for editorial calendars, or even educational assistants that pull the latest lecture recordings into a learning platform. By integrating this server, AI assistants can offer real‑time, personalized video suggestions, automate content ingestion pipelines, or power chatbots that answer “What’s new in my watch‑later list?” queries without any custom backend code.