Transcribe audio to text with OpenAI Whisper running in your browser. 100+ languages. 100% private — never uploads.
AI Audio Transcription
Upload an audio or video file — OpenAI Whisper transcribes it to text right in your browser. 100+ languages with auto-detection. Your audio never leaves your device.
Drag audio/video here or click to upload
MP3, WAV, M4A, OGG, MP4, WebM up to 100 MB. Best results: under 10 minutes per file.
100% Private
Audio never leaves your device. Whisper runs in your browser via WebAssembly.
100+ Languages
Auto-detect or pick the language. Whisper outperforms most commercial STT services.
TXT + SRT Export
Download plain text or SRT subtitle files. Timestamps preserved.
This tool runs OpenAI Whisper — the state-of-the-art open-source speech-recognition model — directly inside your browser using Transformers.js (the JavaScript port of HuggingFace Transformers). Whisper is the same technology used by services like Otter.ai, Rev, and Descript, but those upload your audio to their servers and charge subscription fees. This tool processes everything locally: your audio is decoded to 16 kHz mono, run through the Whisper neural network in WebAssembly, and the resulting transcript stays on your device. Whisper supports 100+ languages with automatic language detection. The output includes both a plain-text transcript and timestamped segments which can be exported as SRT subtitles for video. Use cases: meeting notes, podcast transcription, interview transcription, video subtitles, accessibility captions, voice memo organisation. For best accuracy, use the small model on clear audio; for speed, use tiny on shorter files. The model files are cached in your browser's IndexedDB after first download, so subsequent uses are instant.