YouTube

关于

YouTube is a robust Model Context Protocol (MCP) server engineered to provide AI agents with advanced capabilities for processing YouTube videos. It allows for the comprehensive extraction of video metadata, including title, description, views, and duration, without requiring video download. Furthermore, it delivers smart transcription services featuring an in-memory processing pipeline, precise Voice Activity Detection (VAD) using Silero, and extensive multilingual support across 99 languages, alongside translation capabilities. Optimized for performance with `yt-dlp` and hardware acceleration for Whisper inference, the server also incorporates intelligent file-based caching to streamline operations and prevent redundant processing.

主要功能

• Smart transcription with in-memory processing, VAD, and 99-language support
• Comprehensive metadata extraction for YouTube videos without download
• 33 GitHub stars
• On-demand translation of transcriptions to any supported language
• Optimized performance using `yt-dlp`, hardware acceleration (MPS/CUDA), and parallel processing
• Intelligent file-based caching to prevent redundant processing

使用案例

• Enabling AI agents to retrieve detailed YouTube video information
• Automating the generation of accurate, time-synced transcriptions for YouTube content
• Providing multilingual transcription and translation services for video analysis platforms