FileHero guide
How to transcribe voice and audio to text
A step-by-step guide to transcribe voice and audio to text, with advice on the settings and checks that matter.
Before you start
Use a quiet recording where the speaker is close to the microphone. The first uploaded-file transcription may need to download a browser speech model.
How to do it, step by step
- Choose a recording or upload an audio file. For live recording, allow microphone access when your browser asks.
- Speak clearly or wait for the uploaded audio to finish transcribing. The first local transcription can take longer while its model downloads.
- Read the transcript while listening to uncertain passages. Correct names, dates and numbers before copying the text.
Check the saved file
Review names, dates, amounts and punctuation against the recording. Speech recognition can produce plausible but incorrect words.
Improve an inaccurate audio transcript
Background music, overlapping speech, accents and poor microphones can reduce recognition accuracy. A transcript is not an exact record unless checked. Use a clearer recording or a shorter clip containing one speaker. Correct important words before reusing the text in a document.