Skip to content

Rapport MetaHuman Avatar

A real-time MetaHuman avatar powered by Unreal Engine and pixel streaming. This demo showcases Rapport's cloud-rendered photorealistic avatars with natural conversation capabilities.

Unreal Engine 5Pixel StreamingVoice Interaction

Click Start Demo. Microphone permission is only requested when voice input is enabled in the embed.

How Pixel Streaming Works

1. Cloud Rendering

Unreal Engine renders the MetaHuman on powerful cloud GPUs at high fidelity.

2. Video Encoding

Frames are encoded to H.264/VP9 and streamed via WebRTC to your browser.

3. Low Latency

Audio input is sent back for real-time lip sync and natural conversation.

MetaHuman vs Diffusion-Based Avatars

Unlike diffusion-based avatars (Hedra, Tavus), MetaHumans use traditional graphics pipelines with skeletal rigs and blendshapes. This provides consistent 60 FPS, full artistic control, and deterministic output - but requires more compute and specialized infrastructure for streaming.