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
// EXECUTORS.team_chat for the approval inbox (api/_lib/approvals.js).
//
// A live coordinator trade or standing order is mirrored into the inbox, so the
// owner can approve it from /approvals, a push action, Telegram or email. When
// they do, the inbox claims the row and calls this with it. We execute the
// matching chat step (the runner re-checks that the row's payload hash equals
// the step's and that the plan is unchanged), then resume the run so the next
// specialist moves. The resume is awaited, not fired and forgotten: Cloud Run
// throttles CPU once the response is sent.
import { createRunner } from './runner.js';
import { createSqlStore } from './store.js';
import { createApprovalsBridge } from './approvals-bridge.js';
import { resolveSquad } from './squad.js';
import { loadPrefs } from './prefs.js';
import * as specialists from './specialists.js';
export async function executeTeamChatApproval(row) {
const runner = createRunner({
store: createSqlStore(),
specialists,
bridge: createApprovalsBridge(),
resolveSquad,
loadPrefs,
userId: row.user_id,
});
const outcome = await runner.executeApprovedStep(row.source_ref, { inboxRow: row });
const step = await createSqlStore().getStep(row.source_ref);
if (step) {
try {
await runner.advance(step.run_id);
} catch (e) {
console.warn('[team-chat] resume after inbox approval failed; the chat resumes on its next refresh', e?.message);
}
}
return outcome;
}