Free · On-device Whisper · SRT export
AI subtitles for screen recordings.
Nothing uploaded.
Every "AI captions" tool wants your video on their server. DemoZoom runs Whisper inside your browser tab — transcribe a recording, download the SRT, and the audio never leaves your machine. Free, no sign-up.
Record & transcribe free →How it works
- Record or import. Capture your screen in the browser, or import an existing MP4/MOV/WebM.
- Export your MP4. Zoom, cuts and polish are rendered locally with WebCodecs.
- Click “Generate AI subtitles”. Whisper downloads once (~40 MB), then transcribes on your device — WebGPU-accelerated when available.
- Download the SRT. Upload it to YouTube/LinkedIn or burn it in with any player. The language is detected automatically.
Why on-device matters
Demos of internal tools, customer data on screen, unreleased features — the recordings you subtitle are usually the ones you can't upload to a random SaaS. On-device transcription means NDA-safe captions: the model comes to the audio, not the other way around.
FAQ
How do I add AI subtitles to a screen recording for free?
Record (or import) your video in DemoZoom, export it, then click 'Generate AI subtitles'. Whisper AI transcribes the audio in your browser and gives you an SRT file to download. Free, no account.
Is the audio uploaded to a server for transcription?
No. The Whisper model runs inside your browser tab (WebGPU with a WASM fallback). The only network request is downloading the model itself once (~40 MB) — your audio never leaves your device, which you can verify in DevTools → Network.
Which languages are supported?
Whisper supports 90+ languages and detects the spoken language automatically — English, Spanish, German, French, Russian, Ukrainian, Portuguese, Japanese and many more.
What subtitle format do I get?
Standard SRT with timestamps — accepted by YouTube, LinkedIn, video players and every editor. You can edit the text in any text editor before uploading.
Can it also remove filler words like 'um' and 'uh'?
Yes. The same on-device speech model powers the 'Cut ums' button in the editor: it finds filler words and voiced pauses and cuts them from the export automatically.
Related
Captions without
the cloud.
Record → export → Generate AI subtitles. Free · 90+ languages · SRT download.
Open DemoZoom →