Speech to Text (Audio Transcription)
Turn recordings into text: voice notes, meetings, interviews, lectures and videos. Choose a file or record from your microphone, and the open-source Whisper speech model writes down what was said, in the original language or translated into English. Download the transcript or subtitles.
How to transcribe audio to text
- Choose an audio or video file, or press Record to use your microphone.
- Optionally pick the spoken language, or translate the speech into English.
- Wait while the model transcribes (the first run downloads about 80 MB).
- Copy the text, or download it as .txt or .srt subtitles.
Private transcription
The Whisper base model (OpenAI, Apache-2.0 licence) runs in your browser, so recordings are never uploaded. That makes it suitable for confidential meetings and personal voice notes.
Subtitles included
The transcript comes with timestamps, so you can download an .srt subtitle file for YouTube, Instagram reels or video editors.
Questions
Which audio formats are supported?
MP3, WAV, M4A, OGG, WebM and the audio track of MP4 and MOV videos, wherever your browser can read them.
How long can recordings be?
Up to 60 minutes. Long recordings take a while on phones; a computer is faster.
How accurate is it?
Clear speech in common languages is transcribed well. Background noise, heavy accents and several people talking at once reduce accuracy, so check names and numbers.
Is my recording uploaded?
No. Transcription happens on your device.