Audio & video to text & summary — automatic speech transcription, key point extraction, and speech cleanup
The microphone uses your browser's speech service. In most browsers, your voice goes to the browser's developer, such as Google or Microsoft. To keep your voice on your device, upload the recording as a file: files are recognized on your device.
Or select any audio / video file (OGG, MP3, WAV, FLAC, M4A, MP4, WebM)
Lecture, meeting, or voice recording is processed directly on your device.
The first transcription downloads the selected offline Whisper model once, then it is cached. An uploaded file never leaves your device.
📖 Vocabulary of specific terms and names (optional)
Without it, rare names, titles and professional terms are recognized as similar everyday words: a colleague's name breaks into two common words, a brand turns into a random noun, a lecture term into familiar nonsense. List such words here, separated by commas, and the model will recognize them more accurately. The rest of the speech is unaffected: nothing is dropped or forced to match the list.
Quick set:
20–30 words work better than a long list: the model reads only the beginning of the prompt and drops the rest.
A name or term recognized incorrectly?
Is it free to use?
Yes, this tool is free and needs no sign-up. Your file is processed right in your browser and is not uploaded to a server. There are no watermarks.