CI METAHUMAN Check private access

A local companion, with a face

Intelligence
with presence.

A photoreal companion you can talk with in real time. The avatar renders on your own GPU host; the model and voice stay on infrastructure you control.

Check private access See how it works

The private live stream is running on the CI tailnet. Sign in with CI Portal to connect.

01THE EXPERIENCE

Not another voice
in a blank box.

MetaHuman gives your private AI a face and a voice. Type a message and watch each reply take shape through expression, gaze, speech, and synchronized lip motion.

The current scene pairs a Reallusion character with an Unreal Engine environment. A browser receives the live picture and sound over WebRTC; the GPU host does the rendering.

01Talk naturallyType a message from your browser
02Read the momentGaze and expression follow the exchange
03Hear every replySpeech and lip motion follow each reply
02UNDER THE HOOD

A short path from
your words to a face.

One turn crosses a small, understandable stack. Your browser carries the conversation; the host does the work.

  1. 01YOU

    Speak or type
    from your browser

  2. 02COMPANION

    Language and voice
    run on your network

  3. 03CHARACTER

    Expression and speech
    render in Unreal

  4. 04BACK TO YOU

    Live video and audio
    arrive over WebRTC

CI

Built to be re-aimed. The companion connects to Ollama or an OpenAI-compatible inference endpoint through configuration. Voice can use local speech services, and conversation memory is opt-in.

↗
03YOUR INFRASTRUCTURE

Private by
where it runs.

Rendering and inference are designed to stay on the machines you choose. CI Portal sign-in and Tailscale keep private previews inside the CI network.

Learn about Companion Intelligence
DESIGNED FORYOUR
OWN
MACHINENO SHARED CLOUD RENDER

Private preview

A presence is live.

The private stream is running on the CI tailnet. Sign in with CI Portal to connect.

Check private access