Transcriptor

概要

Transcriptor provides an MCP server (supporting stdio and HTTP/SSE) designed to extract video transcripts and subtitles using `yt-dlp`. It's compatible with MCP hosts like Cursor and can integrate seamlessly into AI workflows. Beyond basic text, it retrieves extensive video metadata, supports various languages, and intelligently paginates large responses. A standout feature is its Whisper fallback, which transcribes video audio when native subtitles are absent, utilizing either local models or the OpenAI API. While primarily an MCP tool, it also includes an optional REST API for broader accessibility.

主な機能

• Fetch cleaned transcripts or raw SRT/VTT subtitles from multiple video platforms.
• Whisper fallback for audio transcription when subtitles are unavailable (local or OpenAI API).
• Retrieve comprehensive video metadata including title, channel, tags, thumbnails, and chapter markers.
• Docker-first deployment for easy local and remote setup, with optional Redis caching for performance.
• 0 GitHub stars
• Paginate large responses for transcripts and raw subtitles to handle extensive content.

ユースケース

• Enhance AI assistants (like Cursor) with direct access to video content for summarization and analysis.
• Automate the collection of video transcripts and metadata for content analysis or archival purposes.
• Integrate video content into custom applications or workflows via its MCP server or optional REST API.