Official KBO API vs third-party APIs: limitations of the public BCE service and why move to a modern API

In brief

The official SPF Economie SOAP API is free and reliable, but it imposes a high technical cost in 2026: no REST/JSON, no webhooks, no UBO, no Peppol, no SLA. For production projects with a client SLA or KYC/AML needs, migrating to a modern third-party API is economically advantageous once you account for the hidden costs.

What no one tells you about the official API

First: the official SPF Economie API is a free public service and a precious one. Without it, the entire Belgian company data ecosystem would not exist. Everything else, including Company Belgium, depends on it.

But in 2026, starting a serious project directly against the official API means imposing on yourself a needless technical and functional burden. Here is precisely why.

What the official API offers

SPF Economie exposes three channels:

KBOPublicSearch — a SOAP service for company search and consultation

EnterpriseRegister — an authenticated SOAP service for accredited actors (notaries, clerks, banks)

BCE Open Data — a daily FTP download of CSV ZIP files (full + delta)

All free, daily-fed. Official source, trusted data.

The 6 concrete limitations

1. SOAP in 2026

KBOPublicSearch is SOAP/XML. Your modern stack (Node.js, Python, Go, Ruby) must:

• Generate a SOAP client (wsdl2java, suds-jurko, gsoap, etc.)

• Parse XML envelopes with namespaces

• Handle SOAP fault codes, not HTTP

• Serialize requests manually

Cost: +3 to 5 days on initial integration, and a permanent cluster of bugs on schema updates.

2. No webhooks

If you want to react to changes (a client becomes inactive, a new director is appointed, accounts are filed), you must poll. For 10,000 monitored clients:

• Hourly polling = 240,000 requests/day

• Detection with average 30-minute delay

• Official quota frequently hit

With a webhook API, you receive one HTTP notification when the event occurs.

3. No enriched UBO

The official API exposes BCE data (company, address, directors, NACE, establishments). But the UBO register is a separate service (MyMinFin), with:

• Separate eID authentication

• No public REST API

• Limitation to manual consultation for most professional accesses

To run a full KYC, you need to cross BCE + UBO. The official API does not do it for you.

4. No Peppol

Verifying if a company is registered on Peppol (to send it an electronic invoice) requires querying a separate directory (Peppol SMP/SML). Not exposed by the official API.

5. No developer support

The official API offers no:

• Maintained OpenAPI/Swagger documentation

• Test sandbox

• Developer chat or email

• Contractual SLA

• Public status (uptime, incidents)

If you have a bug at 3 a.m. on a Friday, you are alone.

6. No business enrichment

The official API exposes raw data. No:

• Data quality score (flagged inconsistencies)

• Computed financial indicators (company size per EU definition)

• Semantic full-text search

• Advanced filtering (by revenue, headcount, etc.)

• Establishment mapping

• Change tracking (timeline)

When the official API remains the right choice

Let's be honest — there are cases where the official API is amply sufficient:

Academic project: students, research, theses

Weekly bulk processing: a Sunday-night script that ingests the full dump and writes to your data warehouse

Very basic internal tool: a script that verifies 5 VAT numbers per day

Backup / source of truth: cross-check a third-party API against the official source for audit

In those cases, paying for a third-party API would be a waste.

When to migrate to a third-party API

Conversely, migrating makes sense as soon as:

• You are in production with a client SLA: your revenue depends on availability

• You do KYC or AML: need for UBO cascade, archiving, traces

• You invoice via Peppol: verification of registration and reception capability

• You track changes: monitoring, alerts, webhooks

• You have volume > 1000 req/day: the official SOAP starts to choke

• Your team lacks SOAP skills: training cost > API cost

• You want OpenAPI / official SDKs: essential for new integrations

The hidden cost of "staying free"

An honest 12-month calculation for a production project:

The gap is systematically in favor of the third-party API once hidden costs are included.

How Company Belgium fills these gaps

For serious projects, Company Belgium provides the bricks missing from the official API:

• Modern REST/JSON with up-to-date OpenAPI doc — see the endpoints reference

• Webhooks: receive a notification as soon as a company changes (new director, accounts filing, corporate purpose change, UBO declaration)

• Automatic UBO cascade, crossed with the UBO register

• Integrated Peppol check

• SDKs Python / Node.js / TypeScript (see the Python tutorial)

• Contractual SLA on paid plans

• Caching, transparent rate limits — see the reference architecture

All this while remaining synchronized with the official BCE (so the legal source of truth is not lost).

Bottom line

The official API remains an essential tool: it is the source. But for most production projects that do more than read VAT numbers, it imposes a hidden cost (SOAP, polling, manual cross-checks) that quickly exceeds the price of a third-party API.

The right reflex: start with the official API if you prototype, then migrate to an enriched API as soon as a real client expects an SLA. It is not an ideological choice, it is an economic calculation.