Three sub-1B open models read a message and each casts a vote. Quorum returns the verdict and a confidence
you can act on — no training, no paid API. Every example is a real test sentence from a public benchmark, so
you can see when the jury gets it right, and when it doesn't.
01 Choose a domain
02 The message
03 Candidate labels
Sentences you type yourself are kept on the demo server so repeats answer instantly.
⚖
Pick a domain and hear the verdict. The three jurors vote, then Quorum
rules — and the dataset tells you whether it was right.