agent-reachlisted

# Agent Reach — Multimedia & Clean Content Extraction ## Overview Agent Reach empowers AI Agents to break through plain HTML web boundaries, enabling direct extraction and synthesis of multimedia content including **YouTube subtitles, Bilibili video transcripts, RSS subscription feeds, and clutter-free web markdown**. ## When to Use & When NOT to Use ### When to Use - Extracting subtitles/transcripts from YouTube or Bilibili videos for summarization. - Querying RSS/Atom feeds for continuous tracking of newsletters, blogs, and podcasts. - Extracting clean, ad-free Markdown from web pages via reader APIs. - Cross-platform technical topic research across video and text sources. ### When NOT to Use - Simple static web downloads where standard `curl` or `requests` is sufficient. - Large video/audio binary file processing (use `ffmpeg` or `media_info` instead). - Platforms requiring active browser UI automation (use CDP browser tools instead). ## Core Workflow & Tool Commands | Target Source | Tool / Command | Output Format | Key Flags | |---|---|---|---| | **YouTube Subtitles** | `yt-dlp --write-auto-subs --sub-lang "zh-Hans,en" --skip-download <URL>` | VTT / SRT Subtitles | `--skip-download` (No heavy video) | | **Bilibili Transcripts** | `bili video <BV_ID> --subtitles` | JSON / Plain Text | Extracts native AI/CC subtitles | | **Clean Web Text** | `curl -s https://r.jina.ai/<URL>` | Clean Markdown | Strips ads, navbars, and boilerplate | | **RSS / Atom Feeds** | `python3