Meta puts holographic avatars into WhatsApp and Zoom calls

Meta announced holographic avatars for video calls at Connect 2026, using the Muse Realtime Avatar model to approximate a caller's facial expressions. Engadget's Karissa Bell, who tried them, wrote that she did not initially realise she was speaking with an avatar.

On Meta Ray-Ban Display frames, setup takes about five minutes in the Meta AI app: the user captures their face from different angles from their phone camera and answers questions about their voice and interests, with prompts to smile, show teeth and repeat phrases with different expressions. On Meta's new VR glasses the avatar is full-body and requires no manual expression training, relying instead on inward-facing cameras and sensors. The avatars work in WhatsApp and Zoom calls. Engadget reports the feature is in "early preview," with early access on Ray-Ban Display coming "this fall." A Ray-Ban Display user can call others as their avatar, but only while wearing the glasses; VR avatars work glasses to glasses only, and the two device types can't call each other.

Why it matters for digital humans: This is likeness capture moving from a studio process to a five-minute phone flow, aimed at telepresence rather than content generation. Two details matter for practitioners: the same model powers both the agent avatar and the user's own hologram, which suggests Meta is consolidating on one real-time face stack; and the avatar is currently tied to the sender wearing Meta's glasses rather than to the calling app, so WhatsApp and Zoom support is narrower than it sounds. Hands-on impressions are a single reviewer's experience under Meta's demo conditions, not a measured result.