About

A small Danish company that finds people in photos, so the law can be answered.

Run from a Danish ApS in Espergærde. We picked one problem worth doing properly — and we are growing the team to do it right.

The name

Why Ansikt.

Naming a face-search tool is harder than it looks. We wanted a word that was honest about what the tool does, without making it sound like a surveillance product.

ansigt /ˈanˌse̝kd/ · Danish, n.

A face. The forward part of a head. By extension: the part of a thing turned toward the world.

From Old Norse ansikt · "what is seen"

Ansikt /ˈanˌse̝kd/ · the company

Originally a placeholder. We tried fancier names. None of them said what the product was. We kept this one because it does.

Spelled with a k · the original Norse form

01 · Origin

Why this, and why here.

i. The conversation 2024 — 2025

Friends with school-age kids kept circling back to the same problem. Photos at school. Photos at sports clubs. Photos on club websites and team pages. Their kids ended up scattered across services that made uploading easy and finding anything later impossible.

Consent had been given, sometimes. Withdrawn, occasionally. Tracked, almost never. When a parent did ask for everything to come down, the people answering had to guess. Not because they didn't care — because the tools didn't exist.

"The law is fine. The tools are missing."

The market had two flavours: enterprise face-recognition platforms designed for surveillance, and consumer photo apps designed for grandparents. Nothing in between for a school secretary, a compliance officer, or a parent with a 30-day clock and seven systems to search.

So we started building it.

ii. Why Denmark Copenhagen · ongoing

Denmark takes data protection seriously and quietly. Datatilsynet, our supervisory authority, has issued some of the EU's most pragmatic GDPR guidance — not the loudest, the clearest.

That same disposition shows up in how Danish companies build software: the small things are taken seriously; nothing is decorative; the marketing is dry on purpose. We wanted to build a company in that idiom.

It also means our jurisdiction question is settled before anyone asks it. Ansikt is a Danish ApS, hosted on EU infrastructure, governed by Danish and EU law. That's not a feature. It's the floor.

iii. The shape we picked 2024 — present

A small kit, drawn for one job. Find a person across every system you own. Show your work. Hand the regulator a PDF that makes sense.

We turned down adjacent product opportunities — live CCTV plug-ins, public-internet face search, marketing analytics. Each would have made the company bigger. Each would have made it worse.

The product is narrow on purpose. We'd rather do this one thing well, for a long time, than do five things that someone, somewhere, eventually regrets shipping.

02 · Principles

How we work, in six lines.

Posted on the wall, in Danish. Translated here for the website.

Boring is a compliment.

A compliance product should be predictable, slow to surprise, and quick to answer. We measure ourselves on uptime and latency, not on novelty.

Det kedelige er det fineste

ii.

Show your work.

Every search and export carries the trail of how we got there. If we can't explain how we got an answer, we won't ship the answer. Auditors and DPOs are the audience.

Vis hvad du har lavet

iii.

Hold less than you have to.

A face thumbnail and a vector per face — the full original image stays where it lives. We don't copy more than we need, and we delete what we don't. Article 5(1)(c), as a habit.

Hold mindre end du må

iv.

Don't sell what people are.

No data brokers. No model-training partnerships. No public-internet face search. Your archive is yours. The vectors derived from it are yours. Full stop.

Sælg ikke hvad folk er

A small kit, kept small.

Every feature is asked to defend itself. If it doesn't survive a quarterly cull, it goes. The roadmap is a deletion exercise as often as an addition one.

Et lille værktøj, holdt lille

vi.

Speak plainly.

No "AI-powered." No "next-gen." No grand verbs. We say what the product does, in the smallest words that fit. The product is interesting enough.

Tal klart

03 · So far

A short timeline. On purpose.

2024 — 2025 The conversation that wouldn't go away.

Friends with school-age kids kept circling the same problem: photos, consent, and where their kids ended up online. The platforms made uploading easy and finding things later impossible.

Q4 · 2025 URL proxy POC.

First sketch of how to serve images that respect a per-face consent record. Not the main product — but the same problem from a different angle.

Christmas · 2025 First main-product prototype.

A holiday week, a stack of test photos, and a question: can we make the search step disappear?

January — March · 2026 Face matching, in earnest.

Different algorithms, different embedding strategies, different distance metrics. Friends pilot the early versions and tell us what does not work.

February · 2026 Consent portal + face blur POC.

A subject-facing portal sketched out. Face blur added to the URL proxy. Both still POC; both still finding their shape.

Now Pre-launch.

Sharpening the product and the pitch. ISO 27001 readiness in progress. Looking for the first paying customers.

If you've read this far

We'd rather have one good conversation than ten signups.

If your organisation handles photos and the GDPR question keeps the DPO up at night, we'd like to hear about it. We answer our own emails.

hej@ansikt.dk Talk to us