Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
{
"id": "assistant",
"contentHash": "3d6fd934ff3071e38098c874ccedd594ec596211ef309ecfbd91f8254fc80934",
"reviewedAt": "2026-09-20T07:28:26.183Z",
"passed": true,
"blockers": [],
"texts": [
"One script tag puts a 3D avatar assistant in the corner of any site. Pick a body, a background, and what it knows about your page. three.ws/assistant",
"The live preview runs in the corner of the page while you configure it, and every change above re-mounts it.",
"Use a built-in body, any three.ws avatar id, a GLB URL, or an agent's body. It answers in chat, or speaks what you type in a speech bubble."
],
"lint": [],
"verification": {
"ok": true,
"checks": [
{
"claim": "The live preview runs in the corner of the page while you configure it, and every change above re-mounts it.",
"kind": "evidence:page",
"target": "https://three.ws/assistant",
"ok": true,
"detail": "live page shows \"Live preview running in the corner. Every change above re-mounts it.\""
},
{
"claim": "Use a built-in body, any three.ws avatar id, a GLB URL, or an agent's body.",
"kind": "evidence:page",
"target": "https://three.ws/assistant",
"ok": true,
"detail": "live page shows \"Use a built-in body, any three.ws avatar id (make one in Forge), a GLB URL, or an agent's body via data-agent.\""
},
{
"kind": "probe:api",
"target": "the one script tag the post promises is live and serves the widget",
"ok": true,
"detail": "HTTP 200, expectations met"
},
{
"kind": "probe:browser",
"target": "the configurator page really mounts its live preview",
"ok": true,
"detail": "2 steps passed"
},
{
"kind": "link",
"target": "https://three.ws/assistant",
"ok": true,
"detail": "HTTP 200"
},
{
"kind": "link",
"target": "https://three.ws",
"ok": true,
"detail": "HTTP 200 via https://three.ws/"
}
]
},
"editor": {
"verdict": "publish",
"scores": {
"accuracy": 5,
"clarity": 5,
"specificity": 4,
"voice": 5,
"professionalism": 5,
"visual": 5
},
"issues": [
{
"severity": "minor",
"area": "specificity",
"quote": "It answers in chat, or speaks what you type in a speech bubble.",
"problem": "The 'answers in chat' half is vague about where answers come from; the live page names a free model chain (Groq, OpenRouter, NVIDIA), which is a stronger, checkable mechanism.",
"fix": "Optional: 'It answers in chat from a free model chain, or speaks what you type out loud in a speech bubble.' Do not sacrifice length or clarity to force it."
},
{
"severity": "minor",
"area": "visual",
"quote": "posts 2 and 3 (thread continuation)",
"problem": "The house contract says every post ships an image; only post 1 carries the frame. For a thread the single hero image on the lead is standard practice, but the rule is stated absolutely.",
"fix": "Accept as is for a thread, or explicitly annotate the pack that media attaches to post 1 only."
}
],
"rewrite": {
"posts": [
"One script tag puts a 3D avatar assistant in the corner of any site. Pick a body, a background, and what it knows about your page. three.ws/assistant",
"The live preview runs in the corner of the page while you configure it, and every change above re-mounts it.",
"Use a built-in body, any three.ws avatar id, a GLB URL, or an agent's body. It answers in chat, or speaks what you type in a speech bubble."
],
"altText": [
"The assistant builder page: a Configure column with avatar, background, mode, name and greeting fields on the left, the one-line script snippet on the right, and the widget's round launcher button in the bottom corner."
]
},
"summary": "Ship it. Every claim matches the live page verbatim, the script tag and preview behave as promised, the screenshot matches the copy exactly, and the voice is clean; the only notes are optional polish on specificity of the model chain and a judgment call on thread media.",
"model": "nvidia:moonshotai/kimi-k3",
"fallbacks": [
"No GCP credentials found. Set GCP_SERVICE_ACCOUNT_JSON, run `gcloud auth application-default login`, or run on GCE/Cloud Run.",
"openrouter 402: {\"error\":{\"message\":\"Insufficient credits. Add more using https://openrouter.ai/settings/credits\",\"code\":402,\"metadata\":{\"limit_source\":\"openrouter_credits\",\"remedy_hint\":\"Add credits at https://openr",
"openai 429: {\n \"error\": {\n \"message\": \"Your account is not active, please check your billing details on our website.\",\n \"type\": \"billing_not_active\",\n \"param\": null,\n \"code\": \"billi"
],
"rewriteFindings": []
},
"editorError": null
}