How to use this tool
How to use
Choose an audio file up to three minutes and 50 MB, select the spoken language, and start. Review the editable transcript, then copy it or save a text file. Split longer recordings before uploading them to this local tool.
When it helps
Useful for short interviews, study notes and excerpts from meetings. Noise, overlapping speakers and proper names can cause mistakes. This lightweight model does not provide speaker identification or guaranteed accuracy.
How it works
A multilingual Whisper model runs inside a Web Worker. This file-based tool does not use remote Web Speech services or request microphone access. Initial model downloads need a connection; inference stays on your device.
Frequently asked questions
- Q. Can I use the microphone?
- A. This tool accepts audio files only and never requests microphone permission. Unsupported browsers show a helpful message.