← Back Home

Speech → Text

Whisper Tiny · runs in your browser · no server

⚠ Runs entirely in your browser

All AI processing happens locally on your device via WebAssembly — no server involved. This is Whisper Tiny, the smallest variant — expect noticeably lower accuracy than Whisper Large or cloud transcription APIs. First load downloads the model (~75 MB). This is a proof-of-concept, not a production service.

Loading Whisper model…
Record or upload audio
Hold to record · click to start/stop

or
📂 Upload audio file (mp3, wav, m4a, ogg…)
Transcript
Transcript appears here…