- Use a ready-made app when it covers your flow. Custom work only pays off for your own rules or systems.
- Check the subscription first: e-conomic Basis has limited API access, and Dinero requires Pro or Total.
- Test the failure cases before you choose. A good integration handles downtime, rounding and retries without duplicate invoices.
If invoices from your own system have to reach e-conomic or Dinero, first check whether a ready-made app from the marketplace can handle your flow. If it cannot, bon.do builds a first version at a fixed price within DKK 15,000–75,000 excl. VAT in 2–6 weeks. The price depends less on the than on what has to happen when something goes wrong.
What does an e-conomic or Dinero integration cost?
A ready-made app costs a monthly subscription set by its vendor. e-conomic itself says that integration prices depend on the vendor (e-conomic). bon.do builds a custom integration at a fixed price on a written scope. A first version costs DKK 15,000–75,000 excl. VAT and takes 2–6 weeks. DKK 15,000 excl. VAT covers one core flow. Integrations cost more, so an accounting integration usually lands above the floor of the range. If it must handle many document types, move history and recover by itself from failures, it belongs in the product range of DKK 75,000–250,000 excl. VAT with a written estimate. Operations and service afterwards cost from DKK 1,500 per month excl. VAT (pricing).
Where in that range an integration lands depends on three things:
- Number of document types. Invoices alone are a smaller project than invoices, credit notes, payments and payouts.
- Number of exceptions. Your own invoice numbers, customer-specific prices and partial refunds need code a standard app does not have.
- How it must fail. An integration that may stall for a day needs less work than one that must recover by itself and never post anything twice.
The accounting system comes on top. According to e-conomic’s price page (15 November 2026), e-conomic costs DKK 249–649 per month excl. VAT for 1,000 entries a year, but full API access starts with the Plus package (e-conomic).
When is a ready-made app enough?
A ready-made app is enough when it covers the whole flow from your system to a posted invoice, including your VAT rules and credit notes. It also has to recover from failures without manual clean-up. e-conomic lists more than 300 apps and integrations, including the large web shop platforms (e-conomic). If an app covers your flow, it almost always costs less than building your own, even after a few years of subscription.
A custom build can pay off when no app connects your own system or handles your VAT rules and exceptions. Typical signs:
- Your own system generates the invoices, and no ready-made app supports it.
- Invoice numbers or VAT rules follow your logic, not the app’s.
- Credit notes and partial refunds must be synced too, and the app only covers invoices.
Several systems in one flow or a documentation requirement do not necessarily call for a custom build. First ask the app vendor to show how your actual flow and its failures are handled.
Our position is simple: buy what exists, and build only the part no app can do. What you do not build, you do not have to maintain.
Which plan gives API access?
Neither e-conomic nor Dinero gives full API access on its entry-level plan. On e-conomic’s Basis, only payroll systems and year-end tools can deliver accounting data through the API, while integrations that only read data are allowed. Full access to integrations comes with Plus, Smart and Komplet (e-conomic). Dinero requires a valid Pro or Total subscription, and a personal integration may make 60 calls per minute (Dinero).
If the Dinero subscription lapses, the integration loses access. Above 60 calls per minute, Dinero answers with status 429. That number matters when history has to move: 1,000 calls take at least 16 minutes and 40 seconds, and one voucher can need several calls. Thousands of old vouchers are a small project with its own plan, not a script that finishes in a minute.
What happens when the integration fails?
The integration has to queue the work, retry by itself and make sure the same invoice is never posted twice. In the clinic platform QualiBooking, which we built, invoices, credit notes and payments go to Dinero. A failed job is retried 8 times over about 3 days before a person takes over. The integration follows four rules:
- Draft first, then post. The document is created as a draft in Dinero and posted afterwards. If a draft is left behind after an unclear failure, the integration reuses that draft instead of creating a new one.
- The amount is checked before posting. The draft total is compared with the platform’s total, and if they differ by more than 5 øre (DKK 0.05), nothing is posted.
- No duplicate documents. Every document gets a fixed key. After an unclear result, the integration checks that key instead of blindly resending, so the same invoice is never created or posted twice.
- Failures are not lost. When the Dinero login expires, the jobs wait until the clinic signs in again. In the end a person can choose to post it by hand, and the job is closed with a note.
The integration sends only name, e-mail, phone number and address to Dinero. CPR numbers and the clinic’s case notes are not included. The accounting part is covered by 123 unit tests and 53 integration tests.
None of those four rules is visible in a demo. They decide whether the bookkeeper has to clean up every month.
What should you ask the supplier?
Ask about failure handling and ownership before you choose an app or a supplier. The 7 questions below show whether the integration survives 1 hour of downtime without lost vouchers or duplicate invoices, and whether the keys are in your own name. Ask for a concrete answer to each before you sign:
- What happens if the accounting system is down for an hour? Is the data sent afterwards?
- Can the same invoice be posted twice if a connection drops in the middle of a call?
- How are VAT and rounding handled, so the total in the books matches down to the last øre?
- Who is told when the login to the accounting system expires?
- Can a person see and fix a failure without calling the supplier?
- What happens to credit notes and partial refunds?
- Are the source code and API keys in your name?
How do e-conomic and Dinero differ technically?
Both belong to Visma, but access works differently. With e-conomic, the integration sends 2 tokens with every call. The first belongs to the app itself. The second is created when the customer approves the app in their e-conomic account. e-conomic has 2 APIs, the older REST API and the newer one e-conomic calls OpenAPI, and both use the same tokens (e-conomic). For integrations across many customers, Dinero uses a sign-in through Visma Connect, where access has to be renewed continuously.
One example from our own work with e-conomic’s API: an entry line without a VAT code is posted without an error and without VAT. No warning. A successful API response is no proof that VAT was applied: compare the posted line and its VAT with what your system intended. Traps like that recur in many accounting APIs, and that is where much of the work lies.
Who owns the integration?
You do, if the contract says so. At bon.do you own the code and the rights from the first commit. Make sure of 2 things in the agreement: the source code is in your own Git , and developer accounts and API keys are created in your company’s name, not the supplier’s. Otherwise your bookkeeping depends on a supplier you cannot replace. Read more in who owns the source code.
How does bon.do approach an integration?
We start with a free project assessment of 1 week, where you get a written scope and an estimate. If a ready-made app from the marketplace can do the job, we say so. If we build, we agree a fixed price for that scope before work starts. We have built both e-conomic integrations and Dinero integrations in and Rust, as part of backends and integrations.
When a system has no API, a software robot can fill in the portal instead. In elmeld, a grid-notification service for electricians, a robot does it in 7 timed steps and tries 5 times before a person takes over.
Key facts
Questions we get asked




