Changes

PR #2073 by sathvikc and ksylvan: feat(youtube): Implement visual text extraction via FFmpeg and OCR

• Implemented FFmpeg and Tesseract-based visual text extraction from YouTube videos, enabling OCR on video frames.
• Added configurable CLI flags for visual extraction parameters, giving users fine-grained control over the feature.
• Fixed support for multi-line yt-dlp outputs and updated syntax compatibility with modern FFmpeg versions.
• Refactored OCR processing to use bounded concurrency, context timeouts, and hardened CLI argument handling for improved stability and security.
• Resolved multiple reliability issues including Tesseract CLI argument handling, racy error handling, and timestamp overflow bugs.

Direct commits

• Docs: make README badges clickable