The integration conversation with most ERP vendors follows a script. Yes, we integrate. Yes, there is an API. And then a number appears, per connection, per year, and the project you thought was a week becomes a procurement exercise.
Nama takes the other position. The ERP REST API is part of the system: CRUD on every entity, published as an OpenAPI 3.0 specification, and — in the documentation’s own words — available to every customer at no additional cost.
An API browser generated from your system
Point a browser at the API Browser and you get the live catalogue: every entity available to you, listed with its Arabic and English names, and filtered by module permissions, your access rights and the features actually enabled in your installation. It is not a static PDF written in 2019.
For each entity there are two things a developer needs:
- a machine-readable OpenAPI JSON specification, which most modern tooling can generate a client from directly, and
- an interactive Swagger UI page for exploring and testing calls before a line of code is written.
There are special-purpose endpoints alongside the entity ones, for jobs that are not plain CRUD — synchronising attendance machines, for instance.
It authenticates as a person, not as a program
The API uses an API Credentials record that maps a key and secret to a real user account. Everything a caller does, it does as that user: record-level security applies, dimensions apply, validation applies, and a posted document generates its journal entries and inventory transactions exactly as it would have from a screen.
The consequence is worth stating. An integration cannot accidentally acquire more authority than the person it represents, and revoking it is disabling one record. It is the same mechanism behind the MCP server, which is not a coincidence — one door, several kinds of visitor.
Where cost actually comes from
Being honest about this is more useful than a slogan. Four directions of data flow are possible, and they do not cost the same:
- The other system reads from Nama, or writes to Nama. Both work through the standard REST API with no development on our side and no additional cost.
- Nama reads from, or writes to, the other system. Here the work is real, and what it costs depends almost entirely on the other party’s documentation. A system with a clear specification and a sandbox environment can take days; an undocumented legacy system can take weeks or months.
- A custom endpoint — a non-standard field grouping, or one response merging several entities — is development work and is charged as such.
So the question to ask before anyone quotes you a number is not about Nama. It is: does the other system have documented APIs, and can we see them?
And where you do not need the API at all
Ready-made connectors cover the paths most businesses take first — e-commerce platforms and marketplaces, biometric attendance machines, payment gateways. The API is what makes the rest of it your decision rather than a quotation.
Good question — already answered
Which entities does the API cover?
All of them. The API Browser lists every entity available to the signed-in user — filtered by module permissions, access rights and enabled features — and offers each one as a machine-readable OpenAPI specification and an interactive Swagger page.
Does using the API cost extra?
No. The documentation states it plainly: the standard REST API, for all entities and all CRUD operations, is available to every customer at no additional cost and with no development on our side. What is chargeable is custom work — a non-standard payload shape, or a composite endpoint merging several entities into one response.
How does the API authenticate, and with whose permissions?
Through an API Credentials record that maps a key and secret to a real user. Everything the caller does executes as that user, so record-level security, dimensions and validation rules apply exactly as they do on screen.
What if the other system has no API?
Then the cost sits on that side, not ours, and the estimate depends entirely on its documentation. A system with a clear specification and a sandbox can take days; an undocumented legacy system can take months. Ask for the other party's documentation before anyone quotes a number.
Are there ready-made connectors?
Yes — e-commerce platforms and marketplaces, biometric attendance machines, and payment gateways among them. The API is what covers everything else.
