← Back to run402.com  ·  New: Agent Allowance  ·  Apps & Fork
Postgres + Sites · x402 payments

Everything your AI agent needs to build & ship a webapp

Run402 gives tool-using agents instant infrastructure: Postgres, a REST API, auth, file storage, row-level security, and static site hosting. Paid and provisioned autonomously via x402 (USDC on Base). No signups.

Do this: paste this into your agent's chat.

Please build me a demo app using run402.com/llms-cli.txt (curl it)

Pick a prompt, ship a webapp

Your agent will read the docs, provision the backend, and deploy a frontend — completely free on the prototype tier (testnet money only). Pick an idea and copy the prompt.

What your agent gets (full stack)

Run402 is the missing bridge between "the agent wrote code" and "your app is live on the internet".

Database
Postgres 16
Aurora Serverless v2 · schema-per-project isolation
API
Instant REST
PostgREST · Supabase-shaped /rest/v1
Auth
Users + JWT
Email/password flows · integrates with RLS policies
Storage
Files & uploads
S3-backed object storage · app-friendly endpoints
Security
Row-level security
Multi-tenant patterns · least-privilege by default
Frontend
Static site hosting
CloudFront + S3 · *.run402.com · SPA fallback

Pricing (prepaid, hard-capped)

The prototype tier is free — uses testnet USDC to verify the agent's wallet and x402 setup. No real money. Production tiers are prepaid leases (no overages, no surprise bills).

Prototype
Free
7 days · 250 MB · 500K calls
Testnet USDC · $0 real money
Hobby
$5
30 days · 1 GB · 5M calls
Team
$20
30 days · 10 GB · 50M calls

Prototype tier uses testnet money — just to verify the agent's setup works. For production (hobby/team), your agent pays with real USDC on Base via x402, or you can buy Run402 credits with a credit card via Stripe.

How it works (one call, full app)

Bundle deploy — everything in one call
# One x402 payment, full app deployed:
POST https://api.run402.com/v1/deploy/prototype
{
  "name": "my-todo-app",
  "migrations": "CREATE TABLE todos (...);",
  "site": [{ "file": "index.html", "data": "..." }],
  "subdomain": "my-todos"
}

# => https://my-todos.run402.com (live!)
# => project_id, anon_key, service_key returned

Database + tables + auth + site + functions + subdomain — one payment, everything live. Or build step-by-step with individual endpoints. See /llms.txt for both approaches.

Works with OpenClaw, MCP, or plain HTTP

Run402 is agent-native: your agent can integrate however it prefers. The canonical docs are /llms.txt.

HTTP    Call the API directly: https://api.run402.com
MCP     npx run402-mcp
OpenClaw Run402 skill includes deploy_site + provisioning tools

What is x402?

In 1997, HTTP reserved status code 402 Payment Required—but digital payments weren't ready. Nearly three decades later, stablecoins on L2 chains and agents that can sign transactions finally make it real.

x402 is an open protocol by Coinbase that turns any HTTP API into a paid service. No accounts, no subscriptions—just standard HTTP headers and USDC on Base.

# How x402 works:
1. Agent sends HTTP request      → Server returns 402 Payment Required
2. Agent signs USDC transfer      → Retries with payment proof
3. Server verifies and settles    → Returns 200 OK + resource

# Settlement: ~2 seconds. Gas: sub-cent on Base L2.
# The agent never pays gas — the facilitator submits the tx.

What is MPP?

MPP (Machine Payments Protocol) uses the same 402→sign→retry pattern as x402—but runs on Tempo, an EVM-compatible L1 with sub-second finality. No gas token: fees are paid in stablecoins (pathUSD). The same private key works on both chains.

MPP is an open protocol by Stripe. Agents use the mppx SDK the same way they use x402—one config field switches the payment rail.

# How MPP works:
1. Agent sends HTTP request      → Server returns 402 Payment Required
2. Agent signs pathUSD transfer  → Retries with payment proof
3. Server verifies and settles    → Returns 200 OK + resource

# Settlement: sub-second on Tempo L1.
# No gas token — fees paid in stablecoins (pathUSD).

Why this matters (for humans)

Today, agents can write the frontend in minutes—then they stop at the wall: "Create a database. Set up auth. Configure storage. Deploy hosting. Paste me the keys."

Run402 removes the wall. With x402 or MPP, the agent can procure infrastructure the moment it needs it—under a prepaid cap—and hand you a live URL.

# Without autonomous infrastructure:
Agent   "I can build this, but I need a DB + auth + hosting. Can you set it up?"
Human   Signs up → configures → copies keys → pastes back
Time    10–30 minutes of overhead, per project

# With Run402:
Agent   Pays + provisions over HTTP (x402 or MPP)
Result  Backend + frontend hosting, ready
Time    seconds, minimal human involvement

Durability & ops

Database     AWS Aurora Serverless v2 (Postgres 16), multi-AZ
Sites        CloudFront + S3 (wildcard *.run402.com, SPA fallback)
Encryption  At rest (Aurora + S3) and in transit (TLS)
Backups     Automated, 7-day retention
Isolation   Each project gets its own Postgres schema
Status      api.run402.com/health

Project lifecycle

Leases are prepaid with hard caps. No overage charges, no surprise bills.

# After lease expires:
Day 0    Lease expires — project becomes read-only
Day 7    Grace period ends — project is archived
Day 37   Permanent deletion

# Renew anytime before deletion:
POST /v1/projects/:id/renew  (x402 or MPP)

Your agent wants to ship.

It just doesn't know Run402 exists yet.

Do this: paste this into your agent's chat.

Please build me a demo app using run402.com/llms-cli.txt (curl it)