What payment terminal software costs

A terminal is not "a screen with buttons". It is a device that takes money from a person with no witness present, and it has to stay honest when the internet drops, the printer runs out of paper and the bill acceptor jams mid-payment. Those cases, not the interface, take most of the work and set the price.

What the work actually covers

The first layer is taking money. Bill acceptor, coin acceptor, card reader, contactless module, QR payment. Each has its own protocol and its own failure modes: a note gets stuck, a card fails to read, the bank answers thirty seconds late. Every one of those has to be handled so that the money does not disappear.

The second layer is working without a connection. A terminal may not stop when the internet drops: it keeps taking payments, stores operations locally and sends them once the link is back. That is separate logic with its own queue and identifiers, so that one payment is never counted twice.

The third layer is accounting. Till reconciliation per shift, how much cash is in the box, how much went through cashless, a daily report. Without it the owner has a terminal but no number.

What drives the price up

The number of hardware types. One bill acceptor is one protocol. A bill acceptor plus a coin acceptor plus a card reader plus acquiring is four separate integrations, each with its own failure mode, and each has to be tested against real hardware rather than an emulator.

Fiscalisation. If the terminal must issue a fiscal receipt, that is separate work: connecting to the fiscal service, handling its outages, queuing receipts while it is unavailable. It is a noticeable part of the project and a bad thing to discover late.

The requirement to run unattended. A terminal with an operator standing next to it and a terminal visited once a week are different systems. The second one needs remote diagnostics, updates without a site visit, and alerts that something went wrong — before the customer notices.

Order of magnitude

For a self-service terminal covering payments, offline operation, till reconciliation and basic reporting, the reference point is roughly EUR 10,000–12,000. That is not a price list but an order of magnitude: the actual figure depends on how many hardware types are involved, whether fiscalisation is required, and how complex the service logic itself is.

Why there is no calculator: two car washes with the same sign outside can run different controllers, different acquiring banks and different reporting requirements. A form that produces a figure from three checkboxes either lies or quotes a ceiling "just in case". Half an hour of conversation and then a number someone can stand behind is the honest version.

What not to save on

The operation log. When a dispute arrives six months later — money went in, service was not delivered — the only way to settle it is to see what the terminal did that minute. A system without a detailed log turns every such dispute into one word against another.

Remote access. Driving to a site to restart a program costs more than anything saved during development — and it will happen every month.

Frequently asked

Can we buy something off the shelf instead?

If your hardware is on the supported list and someone else’s service logic suits you — yes, and it will be cheaper. Development is for cases where the service is non-standard or the hardware is unknown to packaged products.

How long does it take?

The bulk becomes visible within weeks, but the real schedule is set by access to live hardware: without it failure modes cannot be tested, and failure modes are what take the time.

Will the terminal work without internet?

It must — that is a baseline requirement, not an option. Operations queue locally and are sent once the connection returns.