Going live

Build and test on pilot first, then move to production. This page covers both.

Test on pilot

Pilot is the staging environment. It has the same endpoints as production, and nothing on it moves real money. Point at the pilot host with a pilot key, and run the whole flow.

• Quote, create an order, and read it back until it reaches completed.
• Drive an order to failed and expired with a reserved value, and confirm your handler branches on each. See Testing on pilot.
• Create the same order twice with one reference, and confirm you get one order. See Errors and retries.
• Send a request you know is wrong, and branch on the code rather than the message. See Errors and retries.
• For an offramp, resolve the account first and handle account_verification_failed.

Switch to production

Pilot        https://engine-api.pilot.pulseshiga.io
Production   https://engine-api.pulseshiga.io

Confirm before you ship

• Your reference is unique for every order, and stored before each create. See Errors and retries.
• You repeat timeouts, 503, and 500 with the same reference, and never with a changed body.
• You read status first, then branch on the code when a request fails.
• Your production webhook endpoint verifies X-Pulse-Signature against the raw request body before processing an event. See Webhooks.
• Your webhook handler stores event ids durably, acknowledges duplicates, and does not depend on delivery order.
• You handle every terminal status: completed, expired, and failed, and stop reading once an order is terminal. See Order status.
• You read amounts as decimal strings, never as numbers.
• Your logs exclude API keys, signing secrets, NINs, BVNs, and full bank account numbers.