Keyframe Labs is making AI avatars that people can sit across a video call from. The Y Combinator-backed startup, a Robinhood Ventures Fund II holding, builds lifelike AI avatars for video calls, giving agents a face, expressions, and natural movement instead of a text box.
The company’s pitch is that video is the default interface for modern work, and AI participants should fit into that format. A support agent that can appear on screen, hold eye contact, and react to the customer is a different experience from a chatbot in a sidebar. Keyframe’s avatars are designed for sales calls, support sessions, and onboarding where presence changes how the interaction feels.
The technical challenge is timing as much as appearance. Avatars must generate expressions and gestures that match the conversation in real time, without the uncanny lag that breaks the illusion. Keyframe Labs is building the pipeline that keeps those responses fluid enough for live calls, so the avatar feels present rather than pre-recorded.
Keyframe Labs joined the Robinhood Ventures Fund II portfolio, which holds roughly 80 companies financed with $250K SAFEs, skews toward Y Combinator alumni, and debuts on the NYSE on August 13 under RVII at $25 a share.
As organizations push AI deeper into customer-facing roles, the interface layer that makes agents feel human may be what separates the deployments people tolerate from the ones they actually enjoy. Keyframe Labs is betting that the avatar, not the chat window, becomes the default way people meet an AI.
The fund’s position in Keyframe Labs signals confidence that the interface layer of AI is as important as the models underneath. If avatars become a standard part of video meetings, the startup that makes them believable stands to capture recurring value in a market that grows with every new AI deployment. Keyframe Labs is still early, but the direction of travel is clear.
