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
/**
* Oracle — live conviction stream (SSE).
*
* GET /api/oracle/stream?network=mainnet&min_score=0
*
* The ingestion augmentor is a separate process, so this endpoint can't read its
* memory — it DB-polls oracle_conviction for rows scored since a cursor (~2s) and
* fans new/updated verdicts to every connected client. Neon serverless HTTP
* can't hold a LISTEN connection, so polling is the right call (matches the
* sniper/club stream pattern). Events: hello, coin, ping, bye.
*/
import { cors, method, rateLimited } from '../_lib/http.js';
import { limits, clientIp } from '../_lib/rate-limit.js';
import { feedSince } from '../_lib/oracle/store.js';
const NETWORKS = new Set(['mainnet', 'devnet']);
const MAX_DURATION_MS = 90_000;
const PING_INTERVAL_MS = 15_000;
const POLL_INTERVAL_MS = 2_000;
export default async function handleOracleStream(req, res) {
if (cors(req, res, { methods: 'GET,OPTIONS', origins: '*' })) return;
if (!method(req, res, ['GET'])) return;
const rl = await limits.mcpIp(clientIp(req));
if (!rl.success) return rateLimited(res, rl);
const url = new URL(req.url, `http://${req.headers.host || 'x'}`);
const network = NETWORKS.has(url.searchParams.get('network')) ? url.searchParams.get('network') : 'mainnet';
const minScore = Math.max(0, Math.min(100, Number(url.searchParams.get('min_score')) || 0));
res.writeHead(200, {
'Content-Type': 'text/event-stream; charset=utf-8',
'Cache-Control': 'no-cache, no-transform',
Connection: 'keep-alive',
'X-Accel-Buffering': 'no',
});
res.flushHeaders?.();
let active = true;
const send = (event, data) => {
if (!active || res.writableEnded) return;
try { res.write(`event: ${event}\ndata: ${JSON.stringify(data)}\n\n`); } catch { /* client gone */ }
};
send('hello', { ts: Date.now(), network });
// Stream only fresh verdicts; the feed endpoint supplies the initial backlog.
let cursor = new Date().toISOString();
const poll = async () => {
if (!active) return;
try {
const items = await feedSince({ network, sinceIso: cursor, limit: 40 });
for (const it of items) {
if (it.score < minScore) continue;
send('coin', it);
}
if (items.length) cursor = items[items.length - 1].scored_at || cursor;
} catch { /* transient DB error — next tick retries */ }
};
const pollTimer = setInterval(poll, POLL_INTERVAL_MS);
const pingTimer = setInterval(() => send('ping', { ts: Date.now() }), PING_INTERVAL_MS);
const stopTimer = setTimeout(() => { send('bye', { reason: 'rotate' }); cleanup(); }, MAX_DURATION_MS);
function cleanup() {
if (!active) return;
active = false;
clearInterval(pollTimer);
clearInterval(pingTimer);
clearTimeout(stopTimer);
try { res.end(); } catch { /* already closed */ }
}
req.on('close', cleanup);
req.on('error', cleanup);
}