How Webinar to Clips Works
AnyTool turns a webinar recording into a clip pack entirely in the browser. It maps speech/silence via audio RMS, detects slide/scene changes via frame differences, and transcribes locally with a self-hosted Whisper model; chapters are formed by silence gaps and transcript topic shifts and auto-titled from their keywords. Each chapter is cut with the self-hosted ffmpeg engine — lossless stream-copy by default, with optional 9:16 reframe and burned-in captions when re-encoding. Outputs are packaged with a bundled ZIP library: MP4 clips, per-clip SRT/VTT, a YouTube chapter list, branded PNG quote cards and a summary.
- Runs client-side — the recording is never uploaded
- Auto chapters from silence, scenes and local Whisper transcript
- Lossless stream-copy cuts (optional reframe/captions)
- One ZIP: clips, SRT/VTT, chapter list, quote cards, summary
- Self-hosted ffmpeg + Whisper — no CDN
- No watermark, no sign-up, no length cap
