Transcribe voice notes, interviews and audio files for free. The Whisper AI listens right on your device — your audio is never uploaded, and there is nothing to sign up for.
or tap to browse — MP3, WAV, M4A, WebM, OGG
Processed on your device, never uploadedFirst time only — the Whisper speech model (about 40 MB) downloads once from a public model library, then it is cached on this device and works much faster. Your audio itself is never uploaded.
Three steps, all on your phone or computer — nothing is sent anywhere.
Transcription runs entirely in this browser tab using a Whisper model that downloads once from a public model library (Hugging Face). Your recordings and files are never uploaded to any server — close the tab and nothing remains.
Yes — no sign-up, no minutes limit from us, no watermark. The AI runs on your own device, so it costs nothing to run.
English works best with the English model. Pick “Other language / auto-detect” and Whisper will try to detect and transcribe 90+ languages, including Urdu and Hindi — smaller languages and noisy audio may be less accurate with this compact model.
The first use downloads the speech model (about 40 MB) so it can run on your device. After that it is cached by your browser, so later transcriptions start almost immediately. Long recordings still take a while to process — a 10-minute file can take a few minutes on a phone.
MP3, WAV, M4A, WebM, OGG and FLAC all work. Voice notes from WhatsApp (.opus inside a .webm/.ogg container) usually work too — if a file fails, convert it to MP3 or WAV and try again.
No. The audio is decoded and transcribed locally in your browser. Only the AI model files are downloaded, once, from a public CDN.