Transcription language

Daisy doesn't ask you to pick a language before a meeting. The on-device transcription model detects the spoken language on its own. The one thing you control is which model is active — because only a multilingual model can handle languages other than English.

English out of the box

Daisy ships with a small English-only model. It's fast and accurate for English meetings, and it's the model Daisy uses until you choose a different one. If all your meetings are in English, there's nothing to change.

Recording in other languages

To transcribe a non-English meeting, switch to a multilingual model:

1. Open Settings and go to the Providers section.
2. Expand Advanced, then Local transcription model.
3. Pick a model tagged multilingual and click Download if you don't have it yet. Bigger models are more accurate but slower.
4. Click Use to make it the active model (● Active).

The multilingual model takes effect on your next recording and auto-detects the spoken language from the audio. To go back to English, set an English model active again.