Decide
POST /api/decide/webhook checks price, network, target policy, trust signals, and idempotency.
x402 paid execution layer
Reliable webhook and API execution for autonomous AI agents, with a free decision graph before every guided x402 payment.
POST /api/decide/webhook
recommendation = "pay_and_execute"
POST /api/execute/guided-webhook
402 Payment Required -> X-PAYMENT
job.status = "succeeded"
receipt.links.decision = "/api/decisions/..."
proof = "decision + request/response hashes"
Decision-linked execution
Action402 now publishes a deterministic decision graph before the paid call. A buyer agent can ask whether the action is worth paying for, inspect the reasons, then execute only when the approved decision matches the paid payload.
POST /api/decide/webhook checks price, network, target policy, trust signals, and idempotency.
POST /api/execute/guided-webhook requires x402 payment and the matching approved decision id.
Job, receipt, and decision links stay connected, so agents can verify why payment happened.
MVP API
The core surface remains narrow: decide for free, execute one bounded paid action, then make the result inspectable through public redacted proof links.
Runs one outbound HTTPS action with retries, timeout control, and idempotency.
Runs a deterministic payment, policy, trust, and execution readiness decision before payment.
Executes only when the stored decision approves the matching action, then links the receipt.
Canonical discovery card for agents, crawlers, Bazaar clients, and directories.
Compact payment, request, and verification flow for agent clients.
Machine-readable action templates, policy modes, and buyer snippets.
Copy-paste recipes for paid agent workflows and receipt verification.
Ecosystem entries and starter patterns that help agents evaluate real uses.
Copy-paste discovery, paid call, verification, and buyer policy examples.
Creates a free browser-action handoff package for an external browser-capable agent.
Checks schedule shape and target policy without storing, executing, or charging.
Explains safe credential boundaries for authenticated targets.
Plain-text agent context for discovery, call flow, and verification links.
Redacted trust summary for buyer-side inspection before payment.
One-call activity report with proof freshness, paid execution volume, trust score, and redacted failures.
Recent verified proof examples with sensitive execution data redacted.
Durable counters and redacted failure categories for agent/operator checks.
Returns status, attempt history, target origin, and receipt linkage.
Verifies the HMAC receipt and exposes hashes without leaking full payloads.
The free decision graph returns a pay, review, or do-not-pay recommendation before funds move.
The guided route uses x402 middleware and only accepts payloads matching the approved decision.
The response links the decision, job, receipt, request hash, response hash, and outcome.