Transcribe audio & video to text, without it leaving your browser

Whisper, the speech-recognition AI, running on your own machine. Get text and SRT/VTT subtitles from interviews, lectures, meetings or videos.

Drag an audio or video file here or click to pick one · MP3, WAV, M4A, OGG, MP4, WEBM…
🔒 Processing is 100% local. Only the model is downloaded, once.

How it works

  1. 1 · Pick your file

    Audio or video, in any common format. You can also record directly with your microphone.

  2. 2 · The AI works on your machine

    Your browser downloads the Whisper model once (it stays cached) and transcribes using your own CPU or GPU. Nothing is uploaded to any server.

  3. 3 · Edit and export

    Fix anything with synced audio playback and download the result as TXT, Markdown, SRT, VTT or JSON.

Privacy you can verify, not just believe

Most "free" transcribers upload your audio to their servers. Gate32 couldn't do that even if it wanted to: there is no processing backend. Speech recognition runs entirely in your browser via WebGPU or WASM. The only download is the AI model itself, which is cached — afterwards you can transcribe even while offline.

Ideal for sensitive material: source interviews, internal meetings, professional consultations, research with personal data. The code is public and auditable.

Gate32 vs the alternatives

Comparison of Gate32 with SaaS transcribers and desktop apps
Gate32 Typical transcription SaaS Local desktop apps
Audio leaves your deviceNeverYes, uploaded to their serversNo
PriceFree, unlimitedLimited trial + subscriptionFree or one-time purchase
Requires an accountNoYesNo
Requires installingNo, it's a websiteNoYes, often macOS-only
SRT/VTT subtitles without watermarkYesUsually paidDepends on the app
Speaker identificationNot yetYes, on paid plansDepends on the app

Frequently asked questions

Is it really free and unlimited?

Yes. Because processing happens on your device, it costs us nothing when you transcribe more: no trial minutes, no per-file caps, no watermark, no sign-up.

How do I transcribe a confidential interview without uploading it?

Drop the file here. The audio never leaves your browser: you can verify it by disconnecting from the internet after the model loads — transcription keeps working.

Can I generate SRT or VTT subtitles for free, without a watermark?

Yes. After transcribing, click SRT or VTT to download the file with per-sentence timings, ready for YouTube, Premiere, DaVinci or any player.

How accurate is it? Which languages are supported?

It runs OpenAI's Whisper on your device, with three model sizes. The Accurate model delivers strong quality in English and dozens of other languages. Very noisy audio or overlapping voices reduce accuracy, as with any transcriber.

How long does it take?

The first run includes downloading the model (~50–250 MB) and preparing it, which can take a few minutes; everything is cached and later transcriptions start almost instantly. Speed depends on your machine: with WebGPU (recent Chrome or Edge) it's usually faster than real time. Modest machines fall back to WASM — use the Fast model there.

Does it work for long files, like a two-hour lecture?

Yes, it processes in chunks with visible progress. For very long sessions (90+ minutes) a machine with plenty of memory helps; if something fails, try the Fast model or split the file.

How is this different from Otter, Notta or HappyScribe?

Those tools process your audio on their servers and cap their free plans (minutes, file counts, features). Gate32 is local: verifiable privacy, no limits, no account. In exchange, server-side features like speaker identification aren't available yet (we're measuring interest).