Connect gateway →
Payment gateway cascade

Payment gateway cascade. Multi-MID failover for checkout.

A cascade is a routing layer above your gateways. Every transaction attempt is scored on decline signals, timeouts, throttling and MCC restrictions. When the primary MID fails, the next healthy MID captures inside 400ms — buyer sees no decline. This is the pattern high-scale operators use instead of hoping one gateway stays up forever.

See pricing →   How it works →

What a payment gateway cascade actually is

Three ingredients:

  1. A prioritized list of MIDs. Route 1 primary, route 2 backup, route 3 fallback. Each MID has its own gateway API credentials.
  2. A scoring engine per attempt. Every transaction is scored on the response signals: hard decline, soft decline, timeout, 3DS challenge failure, MCC restriction, velocity block. Signals are cheap; running them per attempt is what makes cascade work.
  3. An auto-rotation trigger. If the primary returns a failure signal above threshold, the request is re-issued against route 2 immediately with the same card token. Median rotation time under 400ms — inside the buyer's attention span.

Why a single backup gateway is not a cascade

"Backup gateway" usually means a manual switch — the operator notices the primary is down, logs in, changes a setting, hopes buyers refresh. That works for planned outages. It does not work for the two failure modes that actually cost real money:

A cascade handles both. Per-attempt scoring catches silent degradation because failed attempts trigger rotation. Per-attempt routing catches category throttling because only affected transactions rotate.

What Cascade actually does

1

Wire your MIDs

Bring 2-4 gateway API credentials. Order them by preference (usually cheapest or highest-approval first).

2

Score every attempt

Decline signals, 3DS failures, timeouts, MCC restrictions checked per transaction against your route rules.

3

Rotate under 400ms

Failed primary → re-issue against route 2 with same card token. Buyer stays on checkout page. Order tagged with the route that captured.

Cascade signal, not just cascade retry. A dumb retry is: "primary declined, try secondary." That works for outages, fails on legit declines (buyer's card really did not go through). Cascade uses gateway-provided decline codes to distinguish "route-specific problem" from "card-specific problem" — only the first triggers rotation. The second returns a decline to the buyer with a proper reason.

What we do NOT do

When cascade routing pays off

Frequent questions

What is a payment gateway cascade?

A payment gateway cascade is a checkout layer that routes each transaction attempt through the healthiest gateway in a prioritized list of MIDs. When a route declines, times out or throttles, the next route captures the sale — usually under 400ms — so the buyer never sees an error.

Why not just use a single backup payment gateway?

A single backup is a manual failover — someone still has to detect the outage and swap the gateway. A cascade automates that decision per transaction. And running three or four MIDs means a single MID's slow degradation does not silently kill conversion.

Does cascade routing work with any gateway?

Any gateway with an API can join the cascade. Stripe, NMI, Nuvei, Adyen, Airwallex, Checkout.com, Whop, PXP, Coinbase Commerce, Trustly, MyFatoorah, custom HTTP endpoints.

How fast is the failover?

Median rotation under 400ms. Buyer stays on the same checkout page. No visible decline screen. Order writes back to the underlying storefront (Shopify or WooCommerce) tagged with the route that captured.

Wire a cascade. Stop losing revenue to single-gateway failure.

2-4 MIDs behind an own-domain checkout. Live in 24-48 hours.

See pricing →

High-risk processor → · Shopify Payments alternative → · How it works →