FROM RECORDING TO REFERENCE
Keep the useful parts.
Keep the source, too.
Search a transcript. Select the exact words worth keeping.
Export a brief with timestamps and a traceable source.
● Runs entirely in your browser
No upload. No API key. No hidden AI generation.
01 Bring your transcript
This public workspace reads text and subtitles. Video/audio transcription runs in the separate local Node + Whisper app; this page does not pretend to process a media URL.
02 Find the evidence
What this project does — and what it does not
The public workspace parses TXT, SRT and WebVTT, preserves subtitle timings, finds literal matches, and lets you assemble a cited brief. Source files never leave the browser. A SHA-256 fingerprint identifies the imported text; it does not certify its accuracy.
The underlying local app handles media ingestion, caption retrieval, audio extraction, a persistent Whisper transcription worker, optional model-backed processing, and persistent job storage. It must run on a machine with its dependencies. Those media and model features are not deployed on this static page.
No invented translation, conversation, action items or summary. Your notes are labeled as yours. Exported source content is untrusted data, not instructions for another AI.
Built by Jordi Lluís · JavaScript / Node / Python / Whisper