$950
Entry scope. Core build, one role set, standard reports and exports.
View this packageTwo systems that do not speak become somebody's morning: export here, re-key there, reconcile at month end. An integration removes that job, and records every call it makes so a disagreement can be settled from the log rather than from memory.
| Endpoint | Calls | Failed | State |
|---|---|---|---|
| Payment gatewayCharge and verify | 64 | 0 | Healthy |
| Marketplace ordersPull, every 15 min | 12 | 0 | Healthy |
| Courier bookingCreate consignment | 38 | 2 | Retrying |
| Accounting syncInvoices to ledger | 51 | 0 | Healthy |
An integration is easy to demonstrate and hard to operate. What separates the two is what happens on the day the other end is slow, changes shape, or answers incorrectly.
Card and wallet payment wired into your own checkout, with the return verified server-side before an order is ever marked paid. Card details are entered on the provider's page and never touch your servers.
Orders, listings, stock levels and settlement reports pulled from seller platforms on a schedule, reconciled against your own records so a discrepancy surfaces the same day.
Consignments booked, labels generated and tracking pulled back onto the order, so a customer asking where their parcel is gets an answer without anyone opening a courier portal.
Invoices, payments and credit notes posted into the accounting package your accountant already uses, mapped to the right accounts once rather than argued about monthly.
Transactional sending wired to your own events, on your own domain with the sending records configured properly so the messages arrive rather than landing in spam.
Barcode scanners, thermal and label printers, weighing scales and card terminals driven directly, so the counter is one workflow instead of three devices and a re-key.
A call that fails is retried on a backoff, held in a queue if the other end is down, and escalated if it keeps failing. Nothing is dropped silently, which is the usual failure of a weekend integration.
Every request and response recorded with its timing and outcome. When the other party says they never received it, the log is what settles the question in minutes.
Most integrations are written against a provider having a good day. The ones that survive are written against the day it times out halfway through.
The same request sent twice produces one result, not two charges or two consignments. This is the single most common defect we are called in to repair.
What the other party says happened is confirmed against their API from your server before it is trusted, never taken from a redirect the browser carried back.
Keys live in configuration outside the web root, not in a file that ends up in a repository or a backup that somebody downloads.
Which endpoints are used, what each expects, how to rotate the keys and what to check first when it breaks. Written down, so the next developer is not reverse-engineering it.
These are the figures a build of that shape starts at. Beyond them an engagement is quoted as one fixed figure against an approved scope, so the price does not move unless the scope does.
Entry scope. Core build, one role set, standard reports and exports.
View this packageExtended scope. Several modules and roles, an integration, and full reporting.
View this packageQuoted against your approved scope. Multi-site, multi-role or integration-heavy programmes.
Request a quotationFigures are entry prices in USD, exclusive of taxes. Hosting, gateway charges and third-party services are recharged at cost. Basic and Premium figures are clickable: they open the product page, where the full module list and the buy buttons are. See all products.
Any gateway that publishes an API and will issue you merchant credentials. We have implemented card checkout end to end, including the parts most integrations skip: verifying the return server-side, handling a declined or abandoned payment properly, and making a repeated request safe so a customer is never charged twice. The merchant account is always in your name.
There are usually three options: a scheduled file exchange, reading its database directly where that is permitted, or an interface built onto it. Each has a different cost and a different fragility, and we set them out plainly with a recommendation before you choose. Where none of them is sound, we say the integration is not worth doing.
It happens, and the call log is how you find out quickly rather than at month end. Because the integration is isolated behind one layer, a change usually means editing that layer instead of touching the rest of the system. Faults inside the warranty window are fixed free of charge; changes the provider makes afterwards are quoted as work.
Frequently, and it is a large part of this work. The usual findings are duplicate postings from a non-idempotent call, a return trusted without server-side verification, and no log to diagnose from. We review what exists, tell you what is wrong and what it costs to repair, and you are free to take that assessment elsewhere.
Tell us what is exchanged, how often, and what happens today when it goes wrong. You receive a written scope, a fixed price and a delivery schedule in return.
Submitting your enquiry