Nordlet

Blog

Accounting SaaS in 2026: What It Means and How to Choose One

A guide to what accounting SaaS actually is in 2026, how the three product shapes differ, and how to pick the one that matches how your business creates transactions.

Nordlet Team · · 8 min read

Accounting SaaS means accounting software you reach over the internet, that the vendor runs and updates, and that you pay for as a subscription rather than a licence. The data lives in the vendor's infrastructure, upgrades happen without your involvement, and compliance changes arrive as updates rather than as a project.

That was a useful distinction in 2012, when the alternative was a program installed on an office computer. In 2026 nearly everything is cloud-hosted, so the label no longer tells you much. The question worth asking now is not whether a product is SaaS. It is which of three quite different shapes it takes, because that decides who uses it, how it is priced, and whether it can keep up with your transaction volume.

The three shapes of accounting SaaS

Shape Primary user How data gets in Priced by
App-first accounting A bookkeeper, accountant, or business owner People typing, bank feeds, file imports, some integrations Users, or organisations, per month
API-first (embedded) accounting A developer, on behalf of a product Your software calls the API as events happen Usage, typically requests or transactions
ERP suite A finance team, plus operations Modules across the business feed one database Users per month, plus implementation

App-first accounting is the largest category by install base. The interface is the product; the API, where one exists, is a companion for syncing with other tools. This is the right shape when a person is responsible for the books and the transaction volume is something a person can supervise.

API-first accounting inverts that. The interface is the API, and the application is one client of it. This is the right shape when your own software generates the accounting events — a marketplace settling seller payouts, a SaaS product recognising subscription revenue, an e-commerce platform invoicing across borders. Nordlet sits here: every module is reachable through the API, and the application uses the same endpoints external callers get, with no private ones held back.

ERP suites cover accounting alongside inventory, manufacturing, CRM, human resources, and more. They are bought when the accounting problem is inseparable from the operations problem, and they usually come with an implementation project rather than a signup form.

Most bad software purchases in this category are shape mismatches, not vendor mistakes. An app-first tool bought to serve as a platform's accounting backend will need custom middleware. An API bought to replace a bookkeeper will disappoint the bookkeeper.

What actually changed by 2026

Four shifts should affect a purchase made this year.

Electronic invoicing is becoming mandatory, country by country. Several EU member states already require structured e-invoices for domestic business-to-business trade, and the EU's VAT in the Digital Age package sets the direction: from 1 July 2030, cross-border business-to-business digital reporting moves to e-invoicing, with domestic systems aligning by 1 January 2035. A product that emits only a PDF is on the wrong side of that trend. Structured formats matter — Nordlet renders issued invoices and credit notes as EN 16931 UBL for Peppol BIS 3.0, which is the format most of these mandates build on.

One Stop Shop reporting changes in 2027. From 1 January 2027 the OSS and IOSS schemes gain a new correction mechanism, extra IOSS information, and monthly IOSS listings by member state of consumption. Any tool you buy now should be able to preserve corrections linked to their original transactions, not just totals.

Bank connectivity is standard, not a differentiator. Payment services regulation made consented bank access a commodity. Live feeds are now a baseline expectation; what varies is how well a product matches an incoming payment to the right document, and what it does with the ones it cannot match.

"AI bookkeeping" claims need testing. Automatic categorisation is genuinely useful and genuinely fallible. The question to ask a vendor is not whether the model is accurate but what happens when it is wrong: does the correction leave a traceable record, and can a closed period still be protected?

How the pricing models differ

This is where the shapes separate most sharply, and where a spreadsheet beats a feature list.

Model Typical of Cost grows with Awkward when
Per user per month App-first tools, ERP suites Headcount touching the books Your software, not people, creates the entries
Per organisation per month App-first tools serving accountants Number of legal entities or client companies You need books for hundreds of tenants
Per request or transaction API-first tools Volume your product actually generates Volume is spiky and hard to forecast

A worked example makes the difference concrete. Suppose a marketplace needs books for itself plus 200 seller entities, with roughly 40,000 accounting actions a month in total.

  • Under per-organisation pricing at, say, €30 per organisation per month, that is 201 subscriptions and about €6,000 per month, almost none of it related to how much activity actually occurred.
  • Under metered pricing, the same workload is one account with unlimited companies and 40,000 requests. On Nordlet's published plans that is the €50 per month tier with 30,000 requests included plus 10,000 additional requests at €0.003 each, so about €80 per month.

The numbers in that example are illustrative and the per-organisation figure is a stand-in rather than any specific vendor's price. The structural point holds regardless of the rates: per-seat and per-entity pricing charge you for the shape of your organisation, while metered pricing charges you for the work performed. If people do the work, seats are a fair proxy. If software does the work, they are not.

A checklist for choosing

  1. Name the primary user. A person or a program. Everything else follows from this answer.
  2. Count the entities. One company, a group needing consolidation, or many tenants. Group consolidation and multi-tenant isolation sound similar and are different problems.
  3. List the countries. Then ask, per country, what the product actually produces: rates only, a return calculation, a filing-ready register, or submission to the authority. These are four different levels of support and vendors describe all of them as "supports VAT".
  4. Check the write path. Can everything you need be done through the API, or is some of it interface-only? Interface-only steps become manual steps forever.
  5. Test the failure cases. Retry a request. Try to post into a closed period. Try to edit an issued document. The refusals tell you more than the successes.
  6. Read the export list. Reports as XLSX, PDF, or JSON; tax registers in the format the authority expects; a full journal extract. If your data cannot leave, your switching cost is unbounded.
  7. Price it against your own volumes, not the vendor's example customer.
  8. Ask what happens when the rules change. Who ships the update, on what timeline, and at what cost.

When accounting SaaS is the wrong purchase

Two honest cases.

If you are a sole trader with a handful of invoices a month and a straightforward domestic tax position, a subscription may cost more than it saves. The tooling earns its keep at volume, complexity, or both.

If your requirement is a real-time balance inside a wallet or payments product, and statutory books are a separate concern handled elsewhere, what you need is ledger infrastructure rather than an accounting system. The distinction is covered in our guide to immutable double-entry ledger APIs.

FAQ

What does SaaS mean in accounting?

Software as a service means the vendor hosts and maintains the accounting application, you access it over the internet, and you pay a recurring subscription instead of buying a licence. Updates, including compliance changes, are deployed by the vendor rather than installed by you.

Is accounting SaaS safe for financial data?

It can be, and the question to ask is specific rather than general: where is the data stored, who can access it, how are backups handled, what happens to your data if you leave, and does the product keep an audit record of every change. A vendor that answers those precisely is a better sign than one that answers "bank-level security".

What is the difference between accounting SaaS and an accounting API?

An accounting API is one shape of accounting SaaS, the one where the programmable interface is the product rather than an add-on. A product can be SaaS and have no meaningful API at all.

How much does accounting SaaS cost?

It depends on the pricing model more than the feature set. Per-user tools commonly run from roughly €10 to €70 per user per month; per-entity tools charge for each company; metered tools charge for usage. Nordlet's plans start at €10 per month with 3,000 requests included, and every plan includes every module and the full API. Always model your own volumes, because the same product can be cheap or expensive depending on which of these your business generates most of.

Can accounting SaaS handle EU VAT across several countries?

Some can, at different depths. Ask whether the product only knows rates, whether it computes a return, whether it produces a filing-ready file, and whether it submits anything. Nordlet computes OSS and IOSS returns from your invoices with prior-period corrections, and its per-country VAT rate reference covers the EU-27, but domestic return packs ship for three countries today — Lithuania (FR0600), Germany (UStVA) and Poland (JPK_V7M) — and no return is submitted to an authority on your behalf.

Further reading