If you move serious volume, you are not paying for technology — you are paying a toll that scales with your revenue for a service whose cost does not. Paystack gives operations teams direct access to the same rails, priced flat, through one API that drops into virtually anything you already run — your ERP, your homegrown portal, a nightly batch job, even a spreadsheet.
The gap is not a discount we are extending you. It is the difference between a fee indexed to your invoice size and a fee indexed to the actual cost of an ACH entry.
To be explicit: our card rate is interchange + 2.00% + $0.30. Cards are a percentage here too — the saving comes from moving large invoices onto ACH, not from a cheaper card rate.
The $250 monthly platform fee is fixed and always applies — it is not offset or waived by transaction volume. Published competitor rates as of 2026; your negotiated rates may differ. Card interchange is a pass-through on every platform and is excluded from all three figures — this compares platform margin only.
A percentage made sense when the average ticket was $40. On a $5,000 invoice it is a $145 toll for moving a number between two banks. Same settlement, same reconciliation, flat fee. Keep accepting cards here too — interchange + 2.00% + $0.30 — but stop routing five-figure invoices across them.
You already approve invoices in your ERP. You are paying a second system to re-approve them. Keep the approval workflow you have and call the rail directly when it clears.
Bolt-on payment modules charge for the convenience of living next to your ledger. Your ledger stays exactly where it is — Paystack posts to it over the API.
If treasury is still cutting batch files, the cost is not the fee — it is the two FTEs reconciling exceptions. Returns, NSFs, and settlement arrive as typed events.
There is no plugin to wait for, no supported-platform list to appear on, and no certified-partner queue. Paystack is one REST API with a published OpenAPI schema — which means the integration surface is every system you already run, including the ones nobody sells connectors for.
NetSuite, Sage, QuickBooks, Xero, Dynamics — anything with a scripting layer or an outbound webhook can trigger a payment and post the settlement back.
The homegrown portal, the ops console, the twelve-year-old billing system. If you can add an HTTP call to it, it is integrated.
Fire a collection when a subscription renews, a deal closes, or a work order is marked complete — from whatever system owns that event.
Zapier, Workato, n8n, Make, Retool. Standard REST in, signed webhooks out — no custom connector required to get started.
A nightly Python or SQL job over your AP table can disburse a thousand payments and write back every settlement ID.
An Apps Script bound to the sheet your team already lives in is a legitimate integration. Plenty of volume starts exactly there.
A hosted card-capture page you redirect to or drop in an iframe, branded as you — plus instant bank linking and stored payment methods over the API. Card data never touches your servers.
The identical endpoints exposed over MCP, so an internal model or copilot can execute a payment under the caps you set.
One endpoint collects. One signed webhook confirms. Your own invoice number rides along as correlation_id and comes back untouched — so reconciliation is a join, not a project.
Anything with a shell. Cron jobs, ETL steps, a line in a deploy script.
curl -X POST https://api.getpaystack.com/v1/escrows \
-H "Authorization: Bearer $PAYSTACK_KEY" \
-H "Content-Type: application/json" \
-d '{
"amount_cents": 500000,
"rail": "ach",
"counter_party": "vendor_8812",
"correlation_id": "INV-2026-04417"
}'Paystack does not want to be your ledger. It moves money and hands back a settlement record you post against whatever system already owns the invoice.
A published OpenAPI schema at /openapi.json. Generate a typed client in your language in one command — no SDK lock-in, no proprietary DSL.
Settlement, returns, NSF, and payout events are pushed, signed with HMAC-SHA256, and replayable. No polling loops, no missed reversals.
Collect into a custody wallet you control, hold, then disburse to many recipients. Marketplaces, property managers, and contractors get this without a second vendor.
Every transfer carries a correlation ID you set. Daily settlement scans close the loop and surface anything that did not clear — before your controller finds it.
Per-workspace spend caps, per-counter-party blocklists, immutable audit records, and a single operator kill switch across all pending movement.
The honest version: a first integration is usually a few hundred lines and a week of one engineer’s time — a POST where your current SDK call is, a webhook handler where your current one is, and a settlement record posted to the ledger you already have. That is the entire cost of leaving a percentage behind.
Recurring tenant collections, custody wallets held per property manager, owner disbursements, NSF and ACH-return handling, and daily settlement reconciliation — all through the same API documented above. Not a demo environment: real tenants, real returns, real month-end close.
The pricing is not the risk — the cutover is. So there is no cutover.
Point a slice of volume — one entity, one region, one customer cohort — at Paystack while everything else stays put. Nothing is cut over on faith.
Your existing handlers already understand "payment succeeded" and "payment failed". Map them to the Paystack event names once; the shapes are documented and versioned.
The savings are concentrated in your largest invoices. Moving the top decile of payment size off a percentage typically captures most of the annual delta.
No exclusivity, no minimum term on volume. Keep cards on your current processor forever if that is what the business wants.
Everything on this page is the identical infrastructure our agentic customers build on. The day something in your AP process starts approving invoices on its own — a workflow engine, an internal model, a vendor’s copilot — the spend caps, the reversible settlement window, and the immutable audit record are already sitting underneath it. No second vendor, no re-integration, no new security review.
You do not have to believe anything about 2027 to save money in 2026. You just get to skip the migration when you do.
Book 20 minutes or email it over. We will map it line for line against direct rail pricing and tell you what the delta actually is — including the invoices where staying on cards is the right call. No sales sequence, no discovery call before you get a number.