Offline transcription workspace that replaces Otter.ai / Descript — upload audio/video, auto-transcribe with Web Speech API where available, hand-correct on a waveform, and export SRT / VTT / TXT. Timestamps, speaker labels, burn-in preview. Nothing uploaded.
Audio never leaves your device for manual work and exports — SRT/VTT/TXT are generated in this tab via Blob. Web Speech API (when available) does send audio to the browser vendor’s speech service; we flag this and the transcript is still editable offline afterwards. For fully offline long-form transcription, manual segment editing is the reliable path. Close the tab and the media + transcript are gone.