How to Transcribe Audio Without Uploading It
On-device speech-to-text means your recording never leaves your computer. Here's how it works, and which model to pick for what you're transcribing.
Published July 28, 2026
Why on-device matters
Most transcription tools work by uploading your audio to a server, running it through a cloud speech model, and sending the text back. That's fine for a lot of use cases, and it's a real problem for others — an interview under NDA, a legal or medical conversation, an HR meeting, anything where the recording itself is sensitive. On-device transcription sidesteps the question entirely: there's no server in the loop to trust.
How it works
nah.tools' transcription tool runs OpenAI's Whisper model converted to ONNX format, executed entirely in your browser via WebAssembly. The first time you use it, the model downloads from Hugging Face's CDN — that's the only network request involved. After that, the browser caches it, and every transcription from then on runs locally with no network round-trip for the audio itself.
Which model to pick
- English Fast (~39MB) — English-only, the smallest download, the quickest to process. Good default for a quick English voice memo.
- Multilingual (~73MB) — handles non-English languages and accented speech noticeably better, at a moderate size increase.
- High Accuracy (~237MB) — the largest download, best suited to technical vocabulary, multiple speakers, or audio with background noise, where getting it wrong costs more than the extra download and processing time.
Steps
- Open /audio/transcribe and pick a model based on the table above.
- Drop in your audio or video file — the model downloads once, on first use.
- Wait for processing; larger models and longer recordings take more time, especially on lower-powered devices.
- Export as plain text, SRT/VTT subtitles, or a formatted PDF of notes.
If the recording is noisy or has a wide dynamic range, running it through audio normalization first can make speech more consistent for the model to pick up, before you transcribe.
Frequently asked questions
Does any of my audio get uploaded?
No. The only network request is a one-time download of the Whisper model itself (from Hugging Face), which the browser caches afterward. Your recording is processed locally and never leaves the device.
Which model should I pick?
English Fast (~39MB) for quick English-only transcripts, Multilingual (~73MB) as the default balance for anything non-English or accented, and High Accuracy (~237MB) for technical vocabulary, multiple speakers, or noisy audio where mistakes are costly.
Do I need an internet connection?
Only the first time, to download the model. After that it's cached by the browser and transcription runs fully offline — though the page itself still needs to load, so you're not fully disconnected.
What formats can I export?
Plain text, SRT and VTT subtitle files, and a formatted PDF of notes.