Devices

Give your device a face.

Smart displays, kiosks, mirrors, humanoid robots and exhibits. Wireface gives anything with a screen and a voice a live, expressive face that talks as your AI talks, with no video to record and no avatar to train.

Talk to us Try the face

A traveller at a station information kiosk whose screen shows a glowing wireframe face, talking — Illustration: a station help point with a face.

Ideas

Wherever there's a screen

A face makes a device easier to talk to: you can see that it heard you, that it's thinking, and how it feels about the answer. These are a few of the things you could build. The pictures are illustrations, not products you can buy.

A woman brushing her hair at a bathroom mirror; a smiling face in the corner of the glass is talking to her — Smart mirrors
A face in the glass for the morning: the weather, the day ahead and the news, from someone who looks back at you.
Illustration

A humanoid robot in a hotel lobby, its screen face showing the glowing wireframe face, talking to a guest — Humanoid robots
A screen for a head is enough. The face blinks, glances, looks where you point it and shows what it feels as it answers.
Illustration

A smart display on a kitchen counter, its screen showing the glowing wireframe face — Smart displays and speakers
The assistant on the kitchen counter, with a face that listens, answers and reacts, in any look you give it.
Illustration

Two visitors in a gallery laughing as a painted portrait on a framed screen talks to them — Museums and exhibits
Let a portrait tell its own story. Wireface 2D rigs a painting so it can talk, in the painting's own style.
Illustration

How it runs

A face drawn on the device

What it needs

A screen and a browser engine with WebGL2: a kiosk browser, an embedded WebView, Electron, or the browser on a smart display. The face is plain JavaScript, with no build step and no dependencies.
A voice: whichever AI you connect, such as Gemini Live, OpenAI Realtime, ElevenLabs or your own text-to-speech. The face lip-syncs to its audio as it plays.
A connection for the AI, unless your voice runs on the device too. The face itself needs no server.

What you can change

The look: any photo, painting, render, robot or character, with no training. Change it whenever you like.
The voice and the personality: whatever your AI and its instructions say.
Expressions and moods: happy, curious, concerned, amused, sad, or your own.
Colours, wireframe, teeth and movement: from a glowing mesh to a full skin.

Building blocks

Pick the face that fits your screen

Wireface Core

The 3D face: a wireframe or a skin from any photo, with expressions and lip sync, on a canvas.

Try it Core docs

Wireface 2D

Any flat picture made to talk: a cartoon, anime, pixel art or a painting, in its own style.

Try it Wireface 2D

Wireface for three.js

A face as geometry, or a face on a screen inside a 3D scene: for characters, robots and digital twins.

See the demo three.js

Building a device?

Let's give it a face

Tell us what you're making and what it needs to say. We'll help you work out how a face fits, and which licence covers it.

Talk to us See pricing