YouTube Transcript MCP Documentation

Welcome to the documentation for the YouTube Transcript MCP server. This project provides an MCP (Machine-Callable Program) service for fetching and processing YouTube video transcripts, making them available to Claude and other applications.

Documentation Index

Getting Started

• README - Overview, installation, and usage instructions

Project Status

• Progress Tracker - Current status, completed features, and roadmap
• Project Updates - Major architectural changes and development history

Technical Documentation

• Developer Guide - Comprehensive guide for developers working on the codebase
• Fact-Checking Design - Design document for the fact-checking feature

Files and Components

Core Files

• transcript_lib.py - Core library with transcript processing functionality
• transcript_mcp.py - MCP server implementation exposing tools to Claude
• transcript_segment.py - Transcript segment extraction for fact-checking
• search_api.py - Web search client for fact verification

Test Scripts

• test_transcript.py - Test script for transcript functionality
• test_chapter_markers.py - Test script for chapter markers extraction
• test_statistics.py - Test script for video statistics extraction
• test_fact_checking.py - Test script for fact-checking features

Configuration

• requirements.txt - Python package dependencies
• .gitignore - Git ignore file

Output

• logs/ - Directory for log files with transcript results

Core Features

• Transcript Retrieval: Fetch and format transcripts from YouTube videos
• Chapter Markers: Extract and display chapter markers within transcripts
• Video Metadata: Retrieve comprehensive video information
• Video Statistics: Extract view counts, likes, and upload dates
• Language Support: Access transcripts in multiple languages
• Fact-Checking: Tools to help Claude verify claims from video content

Retired Features

For information about features that were implemented but later removed, see the Project Updates document and the Retired Features section in the Progress Tracker.

Getting Started

To get started with using or developing the YouTube Transcript MCP server, follow these steps:

1. Read the README for installation and setup instructions
2. Check the Progress Tracker to understand the current state of the project
3. For development, refer to the Developer Guide
4. Review Project Updates to understand architectural decisions

Project Vision

The YouTube Transcript MCP server aims to provide a bridge between YouTube's transcript data and Claude, enabling rich conversational interactions around video content. Future enhancements will focus on adding richer transcript processing, additional metadata, and advanced analysis capabilities.

Current Capabilities

• Fetch transcripts from any YouTube video with captions
• Merge transcript segments into readable chunks with timestamps
• Extract video metadata (title, author, description) and statistics
• Include chapter markers in transcripts for better navigation
• Support for multiple languages
• MCP integration with Claude Desktop
• Search for information to verify claims from videos
• Extract specific transcript segments for focused analysis
• Find claims within transcripts and provide context
• Comprehensive logging and testing tools

Recent Updates

• Added fact-checking tools to help Claude verify claims from videos
• Enhanced search integration for finding verification information
• Added transcript segment extraction for focused analysis
• Removed speaker identification functionality due to limited effectiveness
• Added comprehensive documentation about architectural decisions
• Simplified codebase by focusing on core functionality
• Improved project documentation organization

For more detailed information about recent changes, see the Recent Updates section in the Progress Tracker.