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 feed.
*
* GET /api/oracle/feed?network=mainnet&limit=50&min_score=0&tier=strong&category=ai
*
* Serves the materialized oracle_conviction cache (one fast indexed read). On a
* cold cache it opportunistically scores a handful of recent coins straight from
* the data brain (no LLM, DB-only, fast) so the feed is never empty before the
* ingestion augmentor has swept. Also returns the conviction-tier backtest so
* the UI can prove the edge.
*/
import { cors, json, method, wrap, rateLimited } from '../_lib/http.js';
import { limits, clientIp } from '../_lib/rate-limit.js';
import { isDbUnavailableError } from '../_lib/db.js';
import { readFeed, convictionBacktest, scoreCoin } from '../_lib/oracle/store.js';
import { recentMints } from '../_lib/oracle/sources.js';
const NETWORKS = new Set(['mainnet', 'devnet']);
const TIERS = new Set(['prime', 'strong', 'lean', 'watch', 'avoid']);
const CATEGORIES = new Set(['meme', 'tech', 'ai', 'culture', 'community', 'political', 'news', 'animal', 'celebrity', 'utility', 'unknown']);
export default wrap(async (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 p = url.searchParams;
const network = NETWORKS.has(p.get('network')) ? p.get('network') : 'mainnet';
const limit = Math.min(200, Math.max(1, Number(p.get('limit')) || 50));
const minScore = Math.max(0, Math.min(100, Number(p.get('min_score')) || 0));
const tier = TIERS.has(p.get('tier')) ? p.get('tier') : null;
const category = CATEGORIES.has(p.get('category')) ? p.get('category') : null;
let items = await safeFeed({ network, limit, minScore, tier, category });
// Cold-start warm: if the cache is empty, score a few recent brain coins
// (DB-only, no LLM) so the page has something real to render immediately.
if (items.length === 0) {
const mints = await recentMints({ network, limit: 8, sinceSeconds: 6 * 3600 }).catch(() => []);
await Promise.allSettled(mints.map((m) => scoreCoin(m, { network, classify: false, persist: true })));
items = await safeFeed({ network, limit, minScore, tier, category });
}
const backtest = await convictionBacktest({ network }).catch(() => []);
return json(res, 200, {
network,
count: items.length,
items,
backtest,
generated_at: new Date().toISOString(),
}, { 'Cache-Control': 'public, max-age=3, stale-while-revalidate=15' });
});
// A read fault degrades to an empty feed, which the cold-start path above then
// tries to warm. A connectivity failure must NOT take that route: it is not "no
// coins scored yet", and treating it as one made the outage far more expensive
// than the outage itself. Every request answered 200 with an empty feed (so the
// page rendered a dead market as fact), and on the way there it fired eight
// scoreCoin() calls that could not possibly persist. Rethrowing hands it to
// wrap() for the shared 503 + Retry-After and skips the pointless warm attempt.
async function safeFeed(opts) {
try {
return await readFeed(opts);
} catch (err) {
if (isDbUnavailableError(err)) throw err;
return [];
}
}