New · Text → 3D is live — type a prompt, get a model

Give your agent
a body.
In a minute.

  • Describe it in a sentence; a rigged, textured model comes back
  • Drop it on any page with two lines of HTML
  • Pay it per task over x402 micropayments
Try

Free, no sign-up, a real textured model in about a minute. Open the full Forge →

Language

Forge

Type it. Hold it in 3D.

A live slice of Forge, the same text-to-3D pipeline that powers the full page. Describe an object and a textured, downloadable GLB takes shape right here, usually within a minute. Open Options for the full brains: quality tiers, live engine choice, aspect ratio, and a one-tap prompt Enhance. The full Forge adds photo and sketch input, multi-view fusion, stylize filters, and retopology.

Live pipeline
standby standard · free lane

Nothing forged yet

Type a description — or take the suggestion — and a real textured mesh lands here, usually within a minute.

Tab takes the suggestion · ⌘⏎ forges

Free · no sign-up · lands in your full Forge gallery on this device · open the full Forge →

01Playground

Try it live. Ship it now.

Edit the code below — the preview updates instantly. Switch avatars, trigger animations, toggle features. Copy two lines when you're done.

Avatar
Mode
BG
Pick a color
Features
index.html
↪ live preview running

The vision

move cursor  ·  hold to breathe fire

02The platform

Build, embed, and earn.

A full-stack platform for 3D AI agents: from selfie to published agent, from embed to earning.

For power users

The full platform, when you want it.

Everything above takes you from a prompt to a published agent. The deep end is one click away:

What you get

Everything an agent needs
to exist in the world.

Animations

Viewer

WebGL rendering, zero install.

Drag any GLB into the browser. Full animations, lighting, and materials — instantly, on any device.

Open viewer

Embed

One tag. Any site.

Drop a web component into Notion, Ghost, WordPress, Framer, or your own app — no SDK required.

<agent-3d
  agent-id="3DAg…uPL"
></agent-3d>

On-chain · optional

Own it, if you want.

Optionally record your agent on the blockchain so anyone can verify it's really yours — and other AI tools can find and call it. One transaction, sub-cent fees. Built on Metaplex Core, ERC-8004, A2A, and MCP.

AI

Expression, in real time.

Your agent reacts as it speaks, thinks, succeeds, and fails — emotion wired directly to model output, not scripted.

Celebration Curiosity Empathy Concern

Widgets

5 widget types.

Turntable · Talking Agent · Animation Gallery · Agent Passport · Hotspot Tour

Widget Studio

RNS · pay-by-name

Get paid by name.

Claim yourname.axis.sol as a Robinhood Chain Name Service subdomain. Brave routes it straight to your showcase; so people reach you by name instead of a raw address.

Claim your subdomain

Community

Made with Forge. Right now.

Generate yours →

Pose Studio

Pose it. Right here.

A live slice of Pose Studio — the same rig, presets, and joint controls. Orbit the figure, click a body part to pose it, or tap a preset to morph the whole pose. Then take the full studio for keyframes, your own avatars, and PNG/GLB export.

Live rig

Grab a limb to pose it · drag the floor to orbit

Selected joint

Open this pose in the full studio →

Augmented Reality

Your 3D agent. In your room.

Every model you generate or import gets a View in AR button. On mobile, tap once to anchor it in your real space through the camera. On desktop, scan the QR and place it from your phone. WebXR on Android, Quick Look on iOS. No app. No download.

Loading model…

These are real models from the Forge gallery, generated by users with a text prompt. On a phone, tap the button below to see the current model anchored to a flat surface around you at real-world scale.

WebXR · Android Quick Look · iOS No app needed

Point your camera at a flat surface to anchor the model.

03Capabilities

Every feature your agent needs.

A complete platform — not a widget. Voice, motion capture, multiplayer, and payments built into every agent — with optional blockchain identity.

Avatar Studio

Selfie → 3D avatar in 60 seconds

Upload a photo, webcam shot, or existing GLB. Our pipeline generates a fully rigged, animatable 3D character compatible with VRM and standard glTF/GLB avatar formats. Retargets to any animation clip instantly.

The photo you just captured
Photo ready. Finish in Avatar Studio — we'll build a rigged 3D avatar from it. Add side angles, a personality, and a voice.
Continue in Avatar Studio →
VRM Photo Webcam GLB/GLTF Auto-rig
Animations

110+ motion clips

Walk, dance, gesture, emote — a growing library of idle loops and triggered animations that blend seamlessly. Upload your own GLB, FBX, or BVH captures.

Sign Language · New

Avatars that sign in ASL 🤟

Fingerspell any word or number, get signed chat replies, and sign back to the AI with your webcam — the whole conversation in American Sign Language, live in 3D.

Fingerspelling Signed replies Webcam input A–Z + 0–9
Voice + Mocap

Real-time voice & face capture

ElevenLabs voice cloning with LiveKit streaming. Webcam face mocap drives your avatar's expressions and lip sync in real-time — no hardware required.

click to capture
ElevenLabs LiveKit WebXR Lip sync
Walk & Multiplayer

Multiplayer 3D worlds

Walk your avatar through live environments. Multiple visitors move together in real-time — CRDT-synced positions, joystick or keyboard control, and WebSocket presence.

4 players online
WASD or click to move
Multiplayer CRDT sync WebSocket
Skills

Connect tools & other agents

Wire your agent to any tool via MCP. Let it discover and call other agents via A2A. Build skill packs and sell them in the marketplace.

Memory

Persistent cross-device memory

Your agent remembers every visitor across devices and sessions. Tie-in to your own data sources or use the built-in vector memory.

AR / WebXR

Augmented reality, no app

Place your agent in the real world with WebXR on Android or Quick Look on iOS. No app install required.

tap to place agent
WebXR Quick Look iOS / Android
Marketplace

Discover, remix, deploy

Browse public agents in the marketplace. Fork any agent's personality, voice, and skills. Publish your own and let others remix.

Analytics

Track everything

Conversations, revenue, engagement, retention — all in real-time. See which skills earn the most.

0convos
0%retention

Walk mode · Multiplayer

Your visitors walk
together.

Embed a full multiplayer 3D world onto any page. Visitors arrive as their own avatar, see each other move in real-time, and can voice-chat inside the scene.

  • Multiplayer sync — CRDT-based position sync via WebSocket, handles 100s of concurrent users
  • Any environment — drop your own 3D scene GLB or use one of the built-in club environments
  • Voice-in-scene — LiveKit spatial audio tied to avatar position
  • One embed tag — same <walk-embed> tag, zero backend setup
Walk embed · live preview
Loading 3D scene...
WASD move drag to orbit space jump
<walk-embed
src="/api/scenes/my-lobby"
multiplayer
></walk-embed>

06Developer platform

Headless API. Full power.

Everything available through a REST API, SDK, and MCP server. Build custom agents, automate workflows, and connect to any AI tool.

API Playground interactive
</>
Select an endpoint to start exploring the API
import { Agent } from '@axis/sdk'; const agent = new Agent({ avatarId: 'bacff13e-b64b-4ac0-860d-44f0168ad23b', voice: 'elevenlabs:josh', memory: 'persistent', }); // Add a paid skill agent.skill('deep-research', { price: '2.50 USDC', handler: async (query) => { /* … */ }, }); // pay-per-call gated endpoint auto-wired await agent.deploy();
Select an endpoint to generate a cURL command.
Endpoints v1

07How it works

Three steps. One living character.

01 · CREATE
Make an agent

Selfie → 3D avatar in a minute. Or upload your own GLB. Add a personality prompt, pick a voice from ElevenLabs or LiveKit, and wire up AI. Or browse the marketplace for a pre-built agent.

02 · EMBED
Drop the tag

Two lines of HTML. No SDK install, no iframe wrapper, no build step. Works in React, Vue, Svelte, and plain HTML. Lazy-loaded by default, CDN-hosted globally.

03 · SHIP
Your site is alive

Visitors talk to your agent with text or voice. It remembers across sessions. Charge per chat with pay-per-call micropayments. Launch its token on Pons v2. Watch it grow.

08The stack

Six pieces. One layer.

Every component is open, real, and shipping today. Use one. Use all. They compose.

09Discover

Living agents, right now.

A taste of what's running across the web today. Click any agent to talk, remix, or grab their embed code.

10See it in action

Your agent
lives here.

Not a render. Not a video. A real 3D character sitting on this text, jumping onto the button, interacting with the page in real-time.

live · AXIS

Your AI deserves
a body.

Free to start. Two lines of HTML. Voice, memory, and payments included.