Data Lab

Build civic language data together

The Data Lab works with the communities who speak each language to build high-quality civic datasets, written and verified by native speakers, for training AI that understands civic discourse in their own languages.

NATIVELY VERIFIED
TRAINING-READY
OPEN DATA

Languages

Progress in every language

Sentence pairs between English and each language. A pair counts as translated once a volunteer submits it, and as verified once a second speaker approves it. Counts update live.