On-device speech-to-text
Provides speech-to-text with a 29-million-parameter model described as 17 MB in size.
Loading page…
Audio and video / product dossier
Earlybird
Open-weight, on-device speech recognition that predicts words before speakers finish them.
Product brief
What Earlybird does.
Earlybird is a 29M-parameter, 17 MB on-device speech-to-text model that predicts each word correctly about 90 ms before the speaker finishes it, with a confidence on every word. Built for voice agents that need to act sooner. Open weights and code.