What languages does Cutroom AI transcribe?

How transcription works

Cutroom AI uses Whisper, running in your browser, so your audio never uploads. To choose the language, turn on Advanced on the Projects page and pick it under “Spoken language”. Until you pick one, it uses your browser's language if that is on the list, and English if not.

Two voices in one recording are labeled Speaker 1 and Speaker 2 in the transcript.

Captions and the clip finder both read the transcript, so a wrong language setting shows up everywhere. Set it before you import.

Filler words by language

Filler word lists exist for six of the sixteen: English, Spanish, French, German, Italian and Portuguese. In those, hard fillers such as um and uh are removed, and soft ones are tagged for you to check. In the other ten you still get a transcript and captions, and pauses are still cut, but fillers are not picked out on their own. You can add your own words in the Cuts tab with “Add a word”.

Fixing misheard words

Speech models trip on names, brands and jargon in every language. Double-click the word in the transcript and choose “Fix spelling”. Pick “All” to change every copy in the video. These remembered spellings also fix the word in new edits in the same browser.

Captions come out in the language you spoke. Caption translation is turned off.

Whatever the language, a close mic and a quiet room give the speech model its best chance.