A Python library (7,130 stars, MIT license) that retrieves YouTube video transcripts and auto-generated subtitles using YouTube's internal API. No headless browser, no official API key required. Supports transcript translation, proxy configuration, and ships a CLI. Last committed 2026-03-23 (ACTIVE maintenance). No known CVEs; the library switched to defusedxml to mitigate XML injection risks.
Referred by: Slack #Innovation @James via @tom_doerr tweet
No direct impact on the current ODS stack or P0/P1 roadmap. The library is Python-only; ODS services are built in Rust (Actix-web) and TypeScript/Node.js. There is no current use case for YouTube transcript extraction in DocStore, PDF Engine, Notification Hub, or any other service in P0-P2. Could become tangentially relevant if ODS builds a future AI content ingestion pipeline (e.g., a knowledge extraction or media processing service). No integration action recommended at this time.