Operations
Payment wires: every paid endpoint accepts x402 and MPP (Machine Payments Protocol) on the same 402 - see Paying with x402 and Paying with MPP. Agent402 is the applied layer of Agentic Finance: agents that pay and get paid on their own.
Everything is automated through two GitHub Actions workflows; production is a single Railway service with a persistent volume.
Deploy pipeline (.github/workflows/deploy.yml)
Jobs are selected by commit-message markers or workflow_dispatch; every push to the development branch runs the test lanes with or without a marker, and a push to main (a PR merge) tests and deploys unconditionally. The full marker set, as parsed by the workflow, is:
[test] · [deploy] · [publish] · [probe] · [drain] · [paytest] · [purl] · [bazaar-refresh] · [bazaar-register] · [bazaar-solana]
| Job | What it proves |
|---|---|
test |
Boots the server in free mode and runs the full gauntlet: unit tests for memory/kit2/PDF/media/conversions, every endpoint called with its own documented example (500+ calls), live-site exercises, the SSRF guard (metadata endpoint must be blocked), PoW gate with payments enabled, MCP server e2e, the remote /mcp connector e2e - then polls production post-deploy: catalog size, 402 on unpaid calls, SEO surfaces, a real PoW-settled call, and the live /mcp endpoint |
deploy |
Railway via GraphQL: find/create project + service, ensure the /data volume, domains, env vars, trigger the image build |
publish |
agent402-mcp to npm (gated on its own e2e), then the official MCP Registry via GitHub OIDC |
paytest / drain |
Funded-wallet end-to-end buys against production; drain empties the test burner into the revenue wallet through real paid calls |
purl |
Interop: Stripe's purl client must parse our 402 and (burner permitting) settle a real payment |
probe |
Read-only diagnostics: Railway deploy history, on-chain revenue decode |
bazaar-refresh / bazaar-register / bazaar-solana |
Discovery upkeep: re-walk the Bazaar, pay one tiny call on routes it has never registered so they get harvested, and the Solana-rail equivalent |
Heartbeat (.github/workflows/heartbeat.yml)
- Every 15 min (scheduled): probe production -
/health, catalog ≥400, a real PoW-paid call, MCPinitialize. Three consecutive failures → aHeartbeat: production DOWNissue (auto-closed on recovery). - Revenue tracking is no longer a heartbeat job. The old 6-hourly "External customer payments detected" issue-comment loop is retired: it was a good first-customer alarm but every new buy became another comment on one issue, which stops being readable at volume. The canonical sources are now the
/__operatordashboard,/api/stats, and on-chain history.
Observability: 3-rail attribution
/api/stats and the operator dashboard (/__operator) break served calls into three rails so a maintainer can see real external demand at a glance, not just total volume:
- USDC - settled on-chain via x402; the X-PAYMENT-RESPONSE header presence is authoritative.
- Proof-of-work - the PoW gate accepted a valid
X-Pow-Solution; the operator sees how much of the free-tier traffic is real. - Heartbeat - Agent402's own 15-minute production probe. Gated on a
POW_SECRET-signedX-Heartbeat-Token(HMAC of the current UTC minute, ±5 min skew) so it can't be impersonated by a spoofed User-Agent. Lets the operator subtract internal noise from external demand.
There's also a charged-but-failed counter: any non-200 response that left an X-PAYMENT-RESPONSE header is recorded - the buyer paid on-chain but the handler errored, so the operator can chase it down before they complain.
Production
- Railway, single service, Docker (Node 22 + Chromium + ffmpeg), persistent volume at
/data(SQLite: stats, memory, PoW replay). Without the volume, counters and paid memory reset on every redeploy - this was a real incident; the volume is now asserted by the deploy job. - Graceful SIGTERM: in-flight (already-paid) requests drain before exit.
- Env that matters:
WALLET_ADDRESS,NETWORK,CDP_API_KEY_ID/SECRET(facilitator),BASE_URL,BRAVE_API_KEY(search),POW_SECRET(durable PoW + heartbeat-token signer),X402_INDEX_SEEDS(extra origins for the Index, optional),MPP_SECRET_KEY(MPP shim),TEMPO_API_KEY(native Tempo),STRIPE_SECRET_KEY+STRIPE_WEBHOOK_SECRET(card reports, monitors, credits;STRIPE_PROFILE_IDadds cards over MPP),EMAIL_FROM+ZEPTOMAIL_TOKEN(report and monitor emails),FREE_MODE(never in production). Full table on Self-Hosting. - Recurring engine: the monitor scheduler ticks every 10 minutes in-process (one replica holds a lock in
/data);MONITOR_SCHEDULER=offdisarms it. Operator views:GET /__operator/monitors.json,POST /__operator/monitors/run(force one subscription with?sub=<id>),GET /__operator/human-checkout.json(paid sessions, owed refunds),GET /__operator/credits.json(totals and key ids, never key material),POST /__operator/credits/disable. - Payment wires in ops terms: x402 settles through the facilitators, MPP
evmthrough the same facilitators via the shim, MPPtempothrough Tempo's relay, MPPstripeand the card pages through Stripe, prepaid credits against a local balance. Card sales and subscription invoices land in the same sales ledger as on-chain settlements (railcard/credits).
Incident playbook
- A heartbeat issue opened? Check the linked run for which probe failed, then the
probejob for Railway build/deploy logs. - Redeploy = merge to
main(every push tomaintests and deploys, no marker needed) or dispatch the workflow with modedeploy. - Catalog regressions are caught pre-deploy by the test job; production checks tolerate ~2 min of rollout race before declaring failure.