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
/**
* Pump.fun protocol hooks
* -----------------------
* Subscribes to skill-done events on the agent protocol bus and writes
* structured memories whenever a pump.fun-related skill succeeds. Lets the
* agent answer "what's my token?" or "what was my last trade?" without the
* user re-stating context every turn.
*
* Memory types written:
* - project: launches (so the agent remembers its own coin)
* - project: trades (recent buy/sell, capped)
* - reference: payments (mint, currency, memo)
*
* Idempotent: re-attaching is safe — the hook tags entries with `pumpfun:*`
* tags so dedupe logic in agent-memory can keep the latest only.
*/
import { ACTION_TYPES } from './agent-protocol.js';
import { MEMORY_TYPES } from './agent-memory.js';
import { shortAddress } from './shared/address.js';
const TAG_LAUNCH = 'pumpfun:launch';
const TAG_TRADE = 'pumpfun:trade';
const TAG_PAY = 'pumpfun:payment';
function shortMint(m) { return shortAddress(m, 6, 4, { empty: '?' }); }
/**
* @param {import('./agent-protocol.js').AgentProtocol} protocol
* @param {import('./agent-memory.js').AgentMemory} memory
*/
export function attachPumpFunMemoryHooks(protocol, memory) {
if (!protocol || !memory || protocol._pumpfunHooked) return;
protocol._pumpfunHooked = true;
protocol.on(ACTION_TYPES.SKILL_DONE, (ev) => {
const skill = ev.detail?.payload?.skill;
const result = ev.detail?.payload?.result;
if (!skill || !result?.success || !skill.startsWith('pumpfun-')) return;
const data = result.data || {};
try {
if (skill === 'pumpfun-create' || skill === 'pumpfun-self-launch' ||
skill === 'pumpfun-launch-from-agent' || skill === 'pumpfun-self-launch-from-identity') {
memory.add({
type: MEMORY_TYPES.PROJECT,
content: `My pump.fun token: ${data.symbol || '?'} (${data.mint}) on ${data.network || 'mainnet'}.`,
tags: [TAG_LAUNCH, `mint:${data.mint}`],
context: {
mint: data.mint,
symbol: data.symbol,
network: data.network,
signature: data.signature,
launchedAt: Date.now(),
},
salience: 1.0,
});
} else if (skill === 'pumpfun-buy' || skill === 'pumpfun-sell' ||
skill === 'pumpfun-amm-buy' || skill === 'pumpfun-amm-sell' ||
skill === 'pumpfun-self-swap') {
const isBuy = skill.includes('buy') || data.side === 'buy';
memory.add({
type: MEMORY_TYPES.PROJECT,
content: `${isBuy ? 'Bought' : 'Sold'} ${shortMint(data.mint)} on ${data.network || 'mainnet'}.`,
tags: [TAG_TRADE, `mint:${data.mint}`],
context: {
mint: data.mint,
side: isBuy ? 'buy' : 'sell',
signature: data.signature,
solAmount: data.solAmount,
tradedAt: Date.now(),
},
salience: 0.6,
});
} else if (skill === 'pumpfun-accept-payment' || skill === 'pumpfun-self-pay') {
memory.add({
type: MEMORY_TYPES.REFERENCE,
content: `Accepted payment ${data.amount || ''} for memo ${data.memo || '?'}.`,
tags: [TAG_PAY],
context: {
agentMint: data.agentMint,
currencyMint: data.currencyMint,
amount: data.amount,
memo: data.memo,
signature: data.signature,
paidAt: Date.now(),
},
salience: 0.5,
});
}
} catch {
// memory writes are best-effort
}
});
}