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
import { createServer, type IncomingMessage, type ServerResponse } from 'node:http'
import { resolve } from 'node:path'
import { loadOperatorConfig, launchConfigSchema } from '../core/config.js'
import { HoodLauncher } from '../core/launcher.js'
import { AutoScheduler, loadAutoConfig } from '../auto/scheduler.js'
import { ProposalStore } from '../auto/proposals.js'
import { ALL_RAILS, createRail } from '../rails/index.js'
import { clientFromOperatorConfig } from '../core/config.js'
import { HoodLauncherError } from '../errors.js'
const dataDir = process.env.HOOD_LAUNCHER_DATA_DIR ?? resolve(process.cwd(), '.hood-launcher')
const port = Number(process.env.PORT ?? 8787)
const operatorConfig = loadOperatorConfig()
const autoConfig = loadAutoConfig()
const launcher = new HoodLauncher(operatorConfig, dataDir)
const proposals = new ProposalStore(dataDir)
const scheduler = new AutoScheduler(launcher, operatorConfig, autoConfig, proposals)
function send(res: ServerResponse, status: number, body: unknown): void {
const json = JSON.stringify(body, (_k, v) => (typeof v === 'bigint' ? v.toString() : v), 2)
res.writeHead(status, { 'content-type': 'application/json' })
res.end(json)
}
async function readBody(req: IncomingMessage): Promise<unknown> {
const chunks: Buffer[] = []
for await (const chunk of req) chunks.push(chunk as Buffer)
const raw = Buffer.concat(chunks).toString('utf8')
return raw ? JSON.parse(raw) : {}
}
/**
* `hood-launcher`'s HTTP API — the programmatic twin of the CLI, sharing the
* same `HoodLauncher` core.
*
* GET /health
* GET /rails preflight status of every rail
* POST /launch { config, theme?, dryRun? } run the full pipeline
* GET /auto/pending list pending autonomous proposals
* POST /auto/tick run one poll-propose cycle
* POST /auto/approve/:id approve + (if LIVE=1) execute a proposal
* POST /auto/reject/:id
* POST /kill engage the kill switch immediately
*/
const server = createServer(async (req, res) => {
try {
const url = new URL(req.url ?? '/', 'http://localhost')
const path = url.pathname
if (req.method === 'GET' && path === '/health') {
return send(res, 200, {
ok: true,
network: operatorConfig.network,
live: operatorConfig.live,
killed: launcher.killSwitch.isEngaged(),
})
}
if (req.method === 'GET' && path === '/rails') {
const client = clientFromOperatorConfig(operatorConfig)
const probe = { name: 'probe', symbol: 'PROBE', description: '', logoUri: '', socials: {}, initialBuyWei: 0n }
const results = await Promise.all(
ALL_RAILS.map(async (name) => {
const rail = createRail(name)
if (!rail.networks.includes(operatorConfig.network)) {
return { rail: name, available: false, reason: `not deployed on ${operatorConfig.network}` }
}
const preflight = await rail.preflight({ client }, probe)
return { rail: name, available: true, preflight }
}),
)
return send(res, 200, { network: operatorConfig.network, rails: results })
}
if (req.method === 'POST' && path === '/launch') {
const body = (await readBody(req)) as { config: unknown; theme?: string; dryRun?: boolean }
const launchConfig = launchConfigSchema.parse(body.config)
const outcome = await launcher.launch(launchConfig, {
dryRun: body.dryRun ?? false,
...(body.theme ? { theme: body.theme } : {}),
})
return send(res, 200, outcome)
}
if (req.method === 'GET' && path === '/auto/pending') {
return send(res, 200, { proposals: proposals.list('pending') })
}
if (req.method === 'POST' && path === '/auto/tick') {
const proposal = await scheduler.tick()
return send(res, 200, { proposal })
}
const approveMatch = path.match(/^\/auto\/approve\/([a-f0-9]+)$/)
if (req.method === 'POST' && approveMatch) {
const result = await scheduler.approve(approveMatch[1] as string)
return send(res, 200, { proposal: result })
}
const rejectMatch = path.match(/^\/auto\/reject\/([a-f0-9]+)$/)
if (req.method === 'POST' && rejectMatch) {
scheduler.reject(rejectMatch[1] as string)
return send(res, 200, { ok: true })
}
if (req.method === 'POST' && path === '/kill') {
launcher.killSwitch.engage('HTTP POST /kill')
return send(res, 200, { killed: true })
}
send(res, 404, { error: 'not_found', path })
} catch (err) {
const status = err instanceof HoodLauncherError ? 400 : 500
send(res, status, { error: err instanceof Error ? err.message : String(err) })
}
})
server.listen(port, () => {
console.log(`hood-launcher API listening on :${port} (network=${operatorConfig.network} live=${operatorConfig.live})`)
})