Tavus is a San Francisco-based AI research lab building what its founders call "human computing" — lifelike AI humans that can see, hear, and respond through real-time video, voice, and text. Founded in 2020 by CEO Hassaan Raza alongside Quinn Favret and Aaron Foss, Tavus started as an AI video personalization tool for sales and marketing and has since evolved into a full conversational video stack used by enterprises and developers.
In November 2025, Tavus raised a $40M Series B led by CRV, with participation from Scale Venture Partners, Sequoia Capital, Y Combinator, HubSpot Ventures, and Flex Capital, bringing total funding to approximately $64M. Alongside the round, the company launched PALs (Personal Affective Links), AI humans that hold persistent memory, take initiative, and communicate face-to-face with users through their device cameras.
The platform is powered by three in-house foundation models. Phoenix-4 generates lifelike facial expressions, head pose, and emotional cues at conversational latency. Sparrow-1 handles audio understanding, including timing, tone, and intent. Raven-1 interprets visual context such as environments, gestures, and expressions. Together they let Tavus render AI characters that hold real-time, multimodal conversations.
Developers can integrate Tavus through its Conversational Video Interface APIs and SDKs, building AI agents into customer support, sales, healthcare intake, education, and companion apps. Enterprise customers use Tavus for personalized prospect videos, AI front-desk experiences, training simulations, and branded AI hosts. With CRV's backing and a maturing model stack, Tavus is positioning itself as one of the foundational layers for AI characters on the open web.