Human conversation is one of the hardest problems in AI.
Today, we're introducing Sparrow-2, our state-of-the-art, real-time conversational understanding model.
It gives Tavus PALs something most voice AI still lacks: understanding what’s happening in a conversation and
Your first PAL is already built.
Introducing new Templates in PAL Maker: ready-made PALs for onboarding, customer service coaching, candidate interviews, and front desk reception.
Face, voice, personality and skills are wired in. Customize yours and it's live in a few clicks.
Big upgrade to Image → Face.
The biggest change: your Face looks a lot more like you.
Better identity preservation, more natural expressions, sharper skin + color, and much stronger support for anime, cartoons, and other styles.
The gap between the image you upload and the
Introducing Image-to-Replica.
Any face that can exist as a single image, whether it's a brand mascot that has lived in a logo for thirty years, a historical figure, or an AI character designed from a single prompt, can now hold a real conversation.
Sparrow-2 is #1 on TurnBench.
It ranked #1 in end-of-turn detection and #1 in interruption handling. Every model gets one submission, scored blind on a held-out test set, with no retries.
Credit to @sesame for an open, third-party benchmark. This is how the field should measure
