# Transcribe audio — speech to text in your browser, no upload

> Transcribe audio to text and export SRT/VTT subtitles in your browser with on-device Whisper. No upload — the model and your audio stay on your device.

Turn speech into text and export subtitles, using a Whisper model that runs in your browser. Your audio and the model both stay on your device — nothing is uploaded.

On-device transcription is the private way to caption a video, take notes from a meeting, or make a recording searchable — none of which you may want to send to a third-party API. Here Whisper runs in the browser, so sensitive audio never leaves the tab.

You get a timestamped transcript you can read, copy, or export as SRT/VTT subtitles. Accuracy tracks recording quality: a clean, close-mic voice does best, while noise and overlapping speakers make it harder.

Trim the recording to the relevant part first with Trim audio, convert it to a supported format with the Format converter, or clean it up with the Vocal remover before transcribing.

## How to

1. **Add the audio.** Drop a recording onto the tool or browse for it.
2. **Run transcription.** The on-device Whisper model converts speech to timed text; weights cache after the first run.
3. **Review the text.** Read through the transcript and timestamps for accuracy.
4. **Export.** Download the transcript as plain text or as SRT/VTT subtitles.

## FAQ

### What can I export?

Plain text for notes, or SRT and VTT subtitle files with timestamps ready to attach to a video.

### How accurate is it?

Clear speech in a quiet recording transcribes well. Heavy accents, crosstalk, and background noise reduce accuracy, as with any speech model.

### Does the model run in the cloud?

No. Whisper runs in your browser and the weights cache locally, so transcription is private and can work offline after the first load.

### Is my audio uploaded to transcribe it?

No. Both the audio and the model stay on your device; nothing is sent to a server.


## Related tools

- [Trim & fade](https://www.safepaper.app/audio/trim-audio)
- [Format convert](https://www.safepaper.app/audio/convert)
- [Vocal remover](https://www.safepaper.app/audio/vocal-remover)
- [OCR](https://www.safepaper.app/docs/ocr)

---

Canonical HTML: https://www.safepaper.app/audio/transcribe
Markdown: https://www.safepaper.app/audio/transcribe.md

There is no upload endpoint — your files are processed in this browser tab. Open your network tab and check.