How it works
Audio to Text uses Whisper, an open speech-recognition model that understands many languages. The model is downloaded once and then runs in your browser, and your recording is never uploaded. Smaller models are faster, larger ones more accurate.
Step by step
- Open Audio to Text and add an audio file with speech.
- Choose the language and the model size.
- Start the transcription and wait – longer recordings take longer.
- Copy the text or download it.
Tips for accurate transcripts
- Use recordings with little background noise – Voice Cleaner can help.
- Select the language that is spoken.
- Choose a larger model for accents, technical terms or poor audio.
- Always proofread names, numbers and specialist vocabulary.
Subtitles instead of plain text
For videos, timed captions are often more useful than a block of text. Audio to Subtitles creates SRT and VTT files from audio, and Video to Subtitles does the same directly from a video file.
Live dictation
To capture speech as it happens – notes, ideas or a meeting – use Live Speech to Text, which writes along while you speak.