Skip to content

KYC on paper: collect ID, check liveness, run sanctions screen, done. KYC in practice for a Hungarian fintech: which provider is MNB-compatible? What does the audit actually want to see? Why did the SumSub integration fail four months in? This post is the version I wish I'd read before our first fintech KYC project.

What MNB actually audits

Hungarian National Bank (MNB) fintech supervision looks at three things on KYC: the risk-scoring logic (and whether it's documented), the chain of evidence (can you reproduce a given customer's onboarding flow from logs?), and the escalation path (when does a flagged customer stop being self-service?).

MNB doesn't certify KYC providers · they audit how you use one. Picking a 'pre-certified' vendor means nothing if your integration is sloppy.

Provider comparison · 2026

SumSub

Best Hungarian document coverage (old format ID, new eID chip read, magyar nyelv the interface). The UI is solid, the webhook reliability is not · expect to build a reconciliation loop. Pricing scales badly past ~5k onboardings/mo. Our default choice for Hungary-first fintech.

Onfido

Cleaner API, better SLAs, worse Hungarian document coverage (the older ID variant misses more often). Pricing is predictable. Default choice when the core market is EU-wide, not Hungary-only.

Stripe Identity

If you're already Stripe for payments, this is seductive. It works · but the risk-scoring customisation is limited. MNB audit fine on paper; in practice, if you need non-standard escalation rules, you'll outgrow it.

Integration anti-patterns (we ate these)

  1. Trusting the provider's `status: approved` without re-running it on your end. A webhook can be spoofed; always verify with an API call server-side before granting access.
  2. Storing the KYC result as a boolean. MNB wants to know WHY · persist the risk vector (score, flags, documents verified), not just pass/fail.
  3. Assuming document OCR data matches sanctions-screening data. Normalize names · Hungarian characters, diacritics, maiden names all cause hits to miss or false-positive.
  4. No versioning on risk-scoring logic. When MNB asks 'why was this customer approved six months ago', you need the scoring rules that were live that day, not the ones now.

Payment gateway parallel

KYC is coupled to payment onboarding. SimplePay (OTP) is the Hungarian default and has its own merchant KYC layer. Stripe + Barion + Revolut Business handle it differently. For EU SEPA + card, Stripe is the least-friction · for HUF-first consumer, SimplePay integration plus manual Stripe fallback is the pattern we ship most.

If you're pre-launch, talk to MNB's fintech supervision desk before you pick your KYC provider. Their pre-clearance takes 4-6 weeks but saves 6 months of rework if your scoring logic doesn't meet their expectations.

ShareXLinkedIn#
Dezső Mező

By

Dezső Mező

Founder, DField Solutions

I've shipped production products from fintech to creator-tooling · for startups and enterprises, from Budapest to San Francisco.

Keep reading

RELATED PROJECTS

Would rather build together?

Let's talk about your project. 30 minutes, no strings.

Let's talk