Model-Context Protocol (MCP) server that uses yt-dlp to fetch YouTube subtitles and expose them to Claude via MCP.
https://github.com/anaisbetts/mcp-youtubeStop watching entire YouTube videos to extract the information you need. This MCP server pulls video subtitles directly into Claude's context, so you can analyze, summarize, and work with video content as easily as you would with text files.
You're researching a new framework, following a tutorial, or trying to understand a complex technical presentation. The information is locked in a 45-minute video, and you need the key points without sitting through the entire thing. Traditional approaches force you to:
This MCP server bridges YouTube and Claude using yt-dlp to extract subtitles and make them available as structured context. Ask Claude "Summarize this YouTube video" and get instant analysis of the full content.
Instant Technical Research: Pull subtitles from conference talks, tutorials, and technical presentations directly into your Claude conversations. Get summaries, extract code examples, or identify specific implementation details without watching.
Documentation from Video Content: Convert recorded demos, walkthroughs, and video documentation into searchable, referenceable text that integrates with your existing knowledge base.
Time-Compressed Learning: Analyze hours of content in minutes. Perfect for staying current with rapidly evolving tools and frameworks when you can't afford to watch everything.
Architecture Decision Research: You're evaluating microservices patterns and found a 90-minute conference talk from a Netflix engineer. Instead of blocking your afternoon, ask Claude to extract the key architectural principles, trade-offs discussed, and implementation recommendations.
Tutorial Content Extraction: Working through a complex deployment tutorial video? Get Claude to pull out the exact command sequences, configuration files mentioned, and troubleshooting steps without rewinding constantly.
Meeting and Interview Analysis: Process recorded technical interviews or team meetings to extract action items, technical decisions, and implementation details that would otherwise require manual note-taking.
The setup is straightforward and fits into existing MCP workflows:
yt-dlp via your package manager (Homebrew, WinGet, etc.)mcp-installer using @anaisbetts/mcp-youtubeNo API keys, no complex authentication, no additional services to manage. The server handles subtitle extraction and formatting behind the scenes.
Development Workflow Integration: Drop YouTube URLs into your Claude sessions alongside code files, documentation, and other project context. Claude can now reference video content as naturally as it references your codebase.
Research and Learning Pipeline: Build a system where you can queue up technical videos, extract their content, and have Claude create structured summaries or comparison documents across multiple sources.
This isn't about replacing video content—it's about making video content as accessible and queryable as text-based resources in your development toolkit. When information is locked in video format, this server is the key that unlocks it for immediate use in your AI-assisted workflows.