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
/**
* GET /api/x402/mcp-perf: the MCP latency SLA feed the dashboard renders.
*
* Two behaviours here are deliberate and easy to "fix" into something worse:
*
* - A breach is still a 200. The dashboard draws the breach itself, so
* answering non-2xx would make a tool over its p95 indistinguishable from
* the endpoint being down.
* - A missing table (the sweep has never run) is 200 with note
* no_samples_yet, while a real read failure is 503. One of those is a cold
* start, the other is an outage, and they must not read alike.
*
* The window is clamped rather than trusted, because it goes into a rolling
* query: an unbounded one is a free way to make the database do arbitrary work.
*/
import { describe, it, expect, vi, beforeEach } from 'vitest';
let health = null;
let failure = null;
vi.mock('../api/_lib/x402/mcp-latency-sweep.js', () => ({
readPerfHealth: async ({ windowHours }) => {
if (failure) throw failure;
return { ...health, window_hours: windowHours };
},
}));
const { default: handler } = await import('../api/x402/mcp-perf.js');
function mockRes() {
return {
statusCode: 200,
headers: {},
body: null,
headersSent: false,
writableEnded: false,
setHeader(k, v) { this.headers[k.toLowerCase()] = v; },
getHeader(k) { return this.headers[k.toLowerCase()]; },
end(payload) { this.headersSent = true; this.writableEnded = true; this.body = payload; },
get json() { return this.body ? JSON.parse(this.body) : null; },
};
}
const req = (url = '/api/x402/mcp-perf') => ({ method: 'GET', url, headers: {}, socket: {} });
beforeEach(() => {
failure = null;
health = {
ok: true,
healthy: false,
tool_count: 12,
breach_count: 1,
breaches: [{ tool: 'render_avatar', p95_ms: 4200, sla_ms: 3000 }],
tools: [{ tool: 'render_avatar', p50_ms: 900, p95_ms: 4200, p99_ms: 6100, samples: 41 }],
};
});
describe('GET /api/x402/mcp-perf', () => {
it('reports a breached SLA as a 200 the dashboard can render', async () => {
const res = mockRes();
await handler(req(), res);
expect(res.statusCode).toBe(200);
expect(res.json.healthy).toBe(false);
expect(res.json.breach_count).toBe(1);
expect(res.json.breaches[0].tool).toBe('render_avatar');
expect(res.json.generated_at).toBeTruthy();
});
it('clamps the requested window to the supported range', async () => {
const res = mockRes();
await handler(req('/api/x402/mcp-perf?window=100000'), res);
expect(res.json.window_hours).toBe(168);
const res2 = mockRes();
await handler(req('/api/x402/mcp-perf?window=6'), res2);
expect(res2.json.window_hours).toBe(6);
const res3 = mockRes();
await handler(req('/api/x402/mcp-perf?window=not-a-number'), res3);
expect(res3.json.window_hours).toBe(24);
});
it('answers 200 no_samples_yet before the sweep has ever written a row', async () => {
failure = new Error('relation "x402_perf_log" does not exist');
const res = mockRes();
await handler(req(), res);
expect(res.statusCode).toBe(200);
expect(res.json.note).toBe('no_samples_yet');
expect(res.json.tool_count).toBe(0);
expect(res.json.error).toBeUndefined();
});
it('answers 503 on a real read failure', async () => {
failure = new Error('connection terminated unexpectedly');
const res = mockRes();
await handler(req(), res);
expect(res.statusCode).toBe(503);
expect(res.json.ok).toBe(false);
expect(res.json.error).toBe('perf_read_failed');
});
it('refuses a method other than GET', async () => {
const res = mockRes();
await handler({ ...req(), method: 'DELETE' }, res);
expect(res.statusCode).toBe(405);
});
});