Geometry fitted to your photo
The SDK fits facial proportions to your photo and prepares an expression rig, on your device. Export one portable character file; the runtime animates it without fitting again.
Wireface Engine
Give any face life.
Photo-fitted facial geometry, automatic procedural rigging, expressions and real-time lip-sync. Make a character from a photo on your own device, free. Bring it to life in your app.
Drawn live in the browser on WebGL2, and lip-synced to whatever voice your AI speaks with: Claude, Gemini, OpenAI, ElevenLabs or your own. The face here is running in this page.
Hear the lip sync
Where it goes
The same face, wherever your AI needs one
Websites
Wireface Chat: a chatbot your visitors can talk to, with a lip-synced face, a real voice, your knowledge and your tools. One line of code.
/chat
AI agents
Wireface Desktop: a face for Claude Code, OpenCode or Codex that floats over your desktop, does the work through your agent and tells you how it's going.
/desktop
Games and 3D
Faces for three.js characters, players' companions and NPCs, mapped onto the head and lit by your scene, with as many on stage as you like.
/threejs
Any picture
Wireface 2D: anime, a painting, pixel art or a photo, rigged and talking in the browser, in your app or on your page.
/2d
Devices
Smart displays, kiosks, mirrors and humanoid robots: anything with a screen and a voice can have a face that reacts as it talks.
/devices
Testing voice agents
Voice Agent Lab: put your agent on a call with a simulated customer, both with faces, and get it scored, recorded and transcribed.
/simulatorThe free showcase
Meet Wireface Desktop
Experience the engine as a talking face on your desktop. Standalone use is free for everyone, including at work.
ClaudeGemini LiveOpenAI RealtimeElevenLabs
Claude CodeOpenCodeCodexthree.jsAny website
Why a rendered face
Drawn live, so it can be anyone and go anywhere
A Wireface isn't a recording and isn't a model trained on one person. It's a face drawn frame by frame on the device, so you can change who it is in seconds, put it wherever there's a screen, and let it react to what your AI says as it says it.
Lip sync to the real voice
The mouth follows the audio your AI generates as it plays, from a clip, a stream or a live call. Nothing is recorded in advance, and with streamed speech the lips close on m, b and p.
Reacts as it talks
Expressions, blinks, gaze and gestures are driven live, so it can look pleased, puzzled or concerned as the conversation turns, and look wherever you point it.
Any AI behind it
Wireface is the face, not the brain. Put Claude, Gemini, OpenAI, a voice from ElevenLabs or your own model behind it, and change your mind later without changing the face.
Wherever there's a screen
Plain JavaScript on WebGL2, with no build step: a website, an app, a game, a desktop, a kiosk or a device with a display. It can even be geometry in a 3D scene.
Light and private
The face is drawn on the device rather than streamed, so it's quick to start and cheap to run, and making a character keeps your photo in the browser.
Any look
One face, as anyone or anything
A glowing wireframe, a photo of a person, a robot, an oil painting or pixel art: the same rig fits them all, with the same expressions and lip sync. Drop in your own picture and it's wearing it a moment later.
For developers
Attach a face in a few lines
The Wireface SDK is the engine behind everything here: ES modules on WebGL2, with no dependencies. Give it a canvas, a face and a voice: in 3D, on any picture, or on a three.js character.
import { createFace, loadCharacter } from '@wireface/sdk';
const face = createFace(document.querySelector('canvas'));
await face.useCharacter(await loadCharacter('/characters/sarah.wface'));
face.express('amused');
face.listenTo(audioContext, aiVoice); // lips follow your AI, liveGive your product a face
Try the face in your browser and talk to it, then put one on your own website for free. Building a game, an app or a device? Tell us about it.