YouTube Transcript MCP Project Progress Tracker

April 4, 2025 ยท View on GitHub

Completed Features

  • Core Transcript Functionality

    • Extract video ID from YouTube URLs
    • Fetch raw transcript data using youtube-transcript-api
    • Support for language selection
    • Error handling for videos without transcripts
    • Intelligent ~10 second transcript segment merging

  • Metadata Extraction

    • Video title retrieval
    • Author/channel information
    • Video description parsing
    • Thumbnail URL retrieval
    • Hybrid approach combining oEmbed API and HTML parsing
    • Video statistics extraction (views, likes, upload date)

  • Chapter Markers

    • Extract chapter markers from YouTube videos
    • Multiple extraction methods for maximum reliability
    • Integrated chapter display in transcripts
    • Standalone chapter marker retrieval tool
    • Chapter-aware transcript formatting

  • MCP Server Implementation

    • get_transcript tool with metadata integration
    • get_video_metadata tool for metadata-only requests
    • list_transcript_languages tool for language discovery
    • get_chapter_markers tool for chapter retrieval
    • Parameter handling for customization

  • Fact-Checking

    • Web search integration for claim verification
    • Transcript segment extraction for focused analysis
    • Claim identification in transcripts
    • Context-aware search capabilities
    • Structured search results formatting for Claude

  • Testing & Logging

    • Command-line test scripts for transcript and metadata
    • Test script for chapter markers extraction
    • Test script for video statistics
    • Test script for fact-checking features
    • Detailed logging with timestamps
    • JSON and text output formats
    • Organized log file storage

  • Documentation

    • Installation and setup instructions
    • Usage examples for each tool
    • Architectural diagrams (system overview, component structure, sequence)
    • Developer guide with implementation insights
    • Project structure documentation
    • Comprehensive project updates documentation
    • Feature implementation and removal details
    • Fact-checking design documentation

  • Project Structure & Maintenance

    • Flattened directory organization
    • Separated log storage
    • Clean module imports
    • Proper error classes
    • Codebase cleanup and simplification
    • Feature evaluation and pruning of ineffective functionality

Retired Features

  • Speaker Identification
    • Pattern-based speaker detection in transcripts
    • Speaker statistics extraction
    • Speaker-labeled transcript formatting
    • Speaker identification dedicated tool
    • Test scripts for speaker identification

Current Project Status

The YouTube Transcript MCP Server is fully functional with the following capabilities:

  • Fetch transcripts from any YouTube video that has captions
  • Display transcripts with ~10 second merged segments for improved readability
  • Retrieve and include comprehensive video metadata (title, author, description)
  • Extract video statistics (view count, likes, upload date)
  • Identify and display chapter markers within videos
  • Provide fact-checking tools for claim verification
  • Extract transcript segments for focused analysis
  • Find claims within transcripts with fuzzy matching
  • Support multiple languages when available
  • MCP integration with Claude Desktop

All core functionality is implemented and working, with comprehensive documentation in place.

Development Roadmap

gantt
    title YouTube Transcript MCP Development Roadmap
    dateFormat  YYYY-MM-DD
    
    section Completed
    Core Transcript Functionality      :done, core, 2023-10-01, 2023-10-15
    Metadata Extraction               :done, meta, after core, 2023-10-25
    MCP Server Implementation         :done, mcp, after meta, 2023-11-05
    Testing & Logging                 :done, test, after mcp, 2023-11-15
    Documentation                     :done, docs, after test, 2023-11-25
    Project Structure                 :done, struct, after docs, 2023-12-05
    Code Cleanup                      :done, cleanup, after struct, 2023-12-15
    Fact-Checking Features            :done, fact, 2024-04-05, 2024-04-15
    
    section Short-term (1-2 months)
    Caching Layer                     :cache, 2024-05-01, 30d
    Request Error Handling            :error, after cache, 20d
    Unit Tests                        :unit, after error, 30d
    
    section Medium-term (3-6 months)
    Enhanced Transcript Processing    :etrans, after unit, 45d
    Transcript Search Functionality   :search, after etrans, 30d
    Enhanced Metadata                 :emeta, after search, 30d
    
    section Long-term (6+ months)
    Multi-video Processing            :multi, after emeta, 60d
    Transcript Analysis Tools         :analysis, after multi, 90d
    Expanded Media Support            :expand, after analysis, 90d
    Advanced AI Integration           :ai, after expand, 60d

Next Steps

Short-term Improvements

  • Performance Optimizations

    • Add caching layer for frequently accessed videos
    • Implement request timeout handling
    • Add retry logic for transient network errors

  • Enhanced Transcript Processing

    • Add option for paragraph-based segmentation
    • Support for SRT export format

  • Testing & Validation

    • Add unit tests with pytest
    • Create a test suite with sample videos
    • Add CI/CD pipeline for automated testing

Medium-term Features

  • Advanced Search

    • Implement within-transcript search functionality
    • Add timestamp jumping for search results
    • Support for regex-based searches

  • Enhanced Metadata

    • Channel subscriber information
    • Related videos suggestions

  • Transcript Analysis

    • Automatic transcript summarization
    • Keyword extraction
    • Topic modeling

Long-term Vision

  • Multi-video Processing

    • Support for processing playlists
    • Batch transcript fetching
    • Comparative analysis across multiple videos

  • Expanded Media Support

    • Support for other video platforms (Vimeo, etc.)
    • Integration with additional media sources
    • Handling of audio-only content

  • Advanced AI Integration

    • Custom embeddings of transcript content
    • Integration with other LLM-powered analysis
    • Domain-specific knowledge extraction

Technical Debt / Known Issues

  • YouTube's page structure may change, potentially breaking the description extraction
  • No automated retry mechanism for failed requests
  • No caching system implemented yet
  • Limited error handling for edge cases
  • No test coverage measurement
  • Search API requires separate API key configuration

Recent Updates

  • Added fact-checking tools: claim verification search, transcript segment extraction, and claim finding
  • Integrated web search capabilities for verifying claims from videos
  • Enhanced transcript processing with segment extraction and fuzzy matching
  • Added dual display of chapter markers: at the top of transcript output and inline within the transcript
  • Added note for developers to run scripts with python3 instead of making them executable
  • Removed speaker identification functionality due to limited effectiveness and redundancy (see Project Updates)
  • Added comprehensive documentation about architectural decisions and development history
  • Simplified codebase by focusing on core functionality that provides the most value
  • Improved organization of project documentation to better reflect current state

Contributing

If you'd like to contribute to this project, please consider addressing one of the items in the Next Steps section above. Please refer to the Developer Guide for details on the architecture and implementation approach.