Introducing Nova Transcribe — free AI transcription in your browser
We just launched Nova Transcribe, a new web tool. Drop in an audio or video file and an AI model (Whisper) writes out what is said — entirely inside your browser.
What it does
- Transcribes audio and video: works with MP3, WAV, M4A and OGG audio as well as MP4 and WebM video. Formats a browser can't play directly, like WMA, AVI or MKV, are converted automatically when you load them
- 20 languages: pick from 20 major languages including English, Japanese, Chinese and Korean
- Three accuracy levels: standard (recommended), a high-accuracy mode when quality matters more than time, and a light mode when you want speed
- Fix it while you listen: results are listed with timestamps — click a timestamp to play from that point and correct anything you misheard on the spot
- Four export formats: SRT and VTT for subtitles, plain text, and Word (DOCX). You can also copy the whole transcript with one click
How to use it
- Drop in an audio or video file
- Choose the accuracy level and the language, then hit "Transcribe" (the AI model downloads on the first run only — after that it is cached, so there is no wait)
- Play the file back, read along, and fix anything that needs fixing
- Pick an export format and download
No sign-up, no installation — it works the moment you open it.
About your data
Transcription happens entirely in your browser (client-side). Your audio and video are never uploaded to a server — the AI model is downloaded to your device and runs there. That makes it safe for recordings you would rather not hand to a third party, like meetings or interviews.
Every feature is free, with no usage limits.
When it comes in handy
- Turning a meeting recording into a first draft of the minutes
- Adding subtitles (SRT) to your own videos without paying for a service
- Keeping interviews or lectures as text you can search later
Novare Orbis offers other free, browser-based tools as well — see the full list.