step 01→
Tell us the problem
Thirty-minute call. You describe the use case, the data, the regulator you answer to. We come back within a week with a one-page deployment outline — not a sales deck, not a quote, not slides.
Free · no NDA at this stage
Self-hosted or EU serversA suite of AI apps — plus the custom software we build between them, in co‑design with your team, to fit your exact problem. Self‑hosted on your hardware, or on our European cloud.
the suite
where it runs
- Self-hosted
- EU cloud
co-design
Your team and ours,
building it together.
the brief
We need our public servants to use AI safely — without anything ever leaving the building.
real products
built with · Municipality of Delft
6–10 weeks to first users · whitelabelled
Trusted across Dutch government and institutions
01what is Arkintel
A growing suite of apps.Custom systems built on them.All on your terms.
Three things, one company. The suite is a growing library of production AI apps you can deploy this month. The custom builds compose those apps into a system built around your problem — with new modules added where the suite stops. The hosting is your call: your hardware, or our European servers. By design, not by promise.
01→
A suite of AI apps
Production apps, ready today — always growing.
A library of production AI apps your team can deploy this month — multi-tenant, SSO, audit-logged, whitelabelled to run under your name. The suite grows: every general primitive we build for one customer joins it for the next.
See the suite→
02→
Custom systems, built to fit
Composed from the suite. New modules where it stops.
Most Arkintel deployments are custom. We compose the suite into a system that fits your problem — and where the suite stops, we build new modules from scratch. The reusable parts make it cheap. The custom parts make it fit.
- Composed from suite apps + custom modules
- Or designed end-to-end from scratch
- Whitelabelled — runs on your domain
- Integrated with the systems you already run
See real customer builds→
03→
Hosted on your terms
You host it. Or we host it on European servers.
Two ways to run it: on your own hardware, inside your perimeter — or on our European servers under EU law. Same software, same compliance posture either way. The choice is about your constraints, not our capability.
- Self-hosted on your own hardware
- Or fully managed by us on European servers
- Always under European jurisdiction
- Same software — the choice is yours
How hosting works→
02how it works
Every engagement starts with a one-page deployment outline before there’s any contract — free, no NDA, no commitment. Then a small senior team ships to real users from week three. Here’s the actual sequence.
30min
first call
1page
deployment outline
6–10weeks
to production
0
commitment up front
- step 02→
Pick the shape
An app off the suite, a composition of several apps, or a custom build from scratch. We tell you honestly which one fits — sometimes the answer is “just use Chat, you don’t need anything custom”.
App · composed · custom - step 03→
Pick where it runs
Your hardware, your VPC, an air-gapped rack, or our European cloud. Same software either way — the choice is about constraints, not capability. We help you size it.
Self-hosted · VPC · EU cloud · air-gapped - step 04→
We deploy and stay
Typical timeline: 6–10 weeks from kick-off to first production users. Then we stay in the loop — model upgrades, monitoring, evals, training. AI isn’t a one-off delivery.
6–10 weeks to production · ongoing partnership
next step
30 minutes, one page back, zero commitment.
The first conversation is free and takes thirty minutes. Within a week we send a one-page deployment outline — not a sales deck, not a quote, not slides.
03the app suite
Production AI apps,
live today.
Five live products — secure, team-ready, and fully traceable. We can host them for you or you can self-host. Select an app below to see how it works.
Interface · multi-modelopen demo
Chat
Private chat. Configured exactly to your liking.
A self-hosted chat platform that runs entirely in your own building. Designed to guarantee there is no leaking of data or sensitive information, making it the perfect choice for sensitive or regulated industries.
- Runs entirely on your own infrastructure
- Configured exactly to your specific requirements
- No data leakage — safe for sensitive industries
- Everything auditable, nothing egressing without a rule
runsself-hosted·EU cloud
chat.arkintel.com
ready
model
you
Compare quantum and classical encryption — one paragraph, plain English.
Auto
→ picked Llama
— all-rounder, runs on your GPUs
sourcesnist.govwikipedia.orgarxiv.org
ask…
tools available// switch model anytime
05in the field
In the room with
the people we build for.
Every custom Arkintel build starts in rooms like these — close work with the public-sector teams, joint ventures and universities we ship to. Briefs become systems because we sit with the people who’ll actually use the thing.
presentation
01 / 02·Municipality of Delft
Presenting at YES!Delft to other municipalities.
YES!Delft · public-sector partners
Composed fromChatKnowledgeTranscribe
in production
02 / 02·JustAskMomo · TU Delft origin
On the main screen, in front of an aerospace engineering lecture.
TU Delft · aerospace engineering lecture · students using it in the room
Composed fromChatKnowledgeExtract
// close work with the people who’ll actually use it — that’s how every Arkintel build gets its shape.
Bring us a brief→05what partners say
In the words of the
people we ship with.
Joint ventures, municipalities, regulated industries — partners who hold us to a higher bar than any pitch deck ever could.
Schuldhulpje
Luc Hijdra
Founder · Schuldhulpje
Featured story · Schuldhulpje
“Arkintel has been an exceptional partner. Not only bringing the product to life, but playing a key role in refining and elevating the vision behind Schuldhulpje.”
Luc Hijdra
Founder · Schuldhulpje
What we built together
Schuldhulpje
Joint venture · 2024 → now
A joint venture turning unopened letters into a clear path out of debt for residents, volunteers and debt counsellors. Anonymous, free, and built for Dutch municipalities — with the AI platform designed and run by Arkintel.
06where it runs · sovereign by design
Your hardware,
or our European cloud.
Two places, same software. European jurisdiction, open-weight models, default-deny egress at the network — either way.
Self-hosted. Inside your perimeter.
Deploy the full stack inside your own data centre, your private cloud account, or an air-gapped rack. We install it, you run it — with our team on the other end of a support channel. Default-deny egress at the network layer, not the application. Air-gap optional, with signed offline update bundles when you choose it.
- On-prem, private cloud, or air-gapped
- Default-deny egress · enforced at network layer
- Open weights — yours to keep
Arkintel-managed. EU-resident. Operated by us.
Deployed on Hetzner’s EU regions — Falkenstein, Helsinki, Nuremberg — a German-headquartered provider, operated end-to-end by Arkintel as a Dutch entity. Outside the reach of the US CLOUD Act and any foreign subpoena. Dedicated tenancy, EU-only egress, same APIs and audit story as the self-hosted path.
- EU jurisdiction · no US exposure
- Hetzner EU regions · operated by Arkintel
- Dedicated tenancy · EU-only egress
Want the best frontier models too?Add Privacy Gate to your self-hosted stack
Need the deployment story in detail?Read the self-hosted deep-dive
07who we work with
Built for teams whose
data can’t leave the building.
Four audiences, same underlying concern: you need AI, but your regulator, your clients or your security team won’t accept “just send it to a US cloud”. We meet you where the regulator does.
Government→
Municipalities, ministries, agencies
Sovereign AI for the public sector — built for citizen data and Dutch procurement reality. On your hardware if you have a data centre, on our European cloud if you don’t. GDPR by architecture, never on a US provider.
Live today
DelftGPT — running inside the City of Delft→
Regulated industry→
Healthcare · legal · finance
AI for environments where “send it to OpenAI” ends the conversation — patient data, case files, contracts, transaction records. Open-weight models, default-deny egress, full audit trail. Aligned with GDPR and the EU AI Act, by architecture.
Typical brief
“We can’t send patient data to OpenAI.”
Enterprise IT→
CIOs and security teams
Give your teams the AI they keep asking for — without handing data to a US cloud. Default-deny egress at the network layer, full audit log, SAML / OIDC via Keycloak into the IdP you already run. Your security team leaves the first call with what they asked for, not a 60-day vendor review.
Typical brief
“Legal vetoed the SaaS vendor. Now what?”
10about Arkintel
A senior engineering team in Delft,
building secure AI for European organizations.
Arkintel is a senior team — engineers, designers, operators — based in Delft, NL. Born inside TU Delft and incubated at YES!Delft. The people designing your system are the people you’ll talk to. No handoffs, no unnecessary layers in the middle, no loss in translation.
- based in
- Delft, NL
- incubated at
- YES!Delft
- building since
- 2016
let’s talk
Want AI you can actually
deploy inside your organisation?
Three doors. Pick one — we usually reply within one working day.
08common questions
Procurement-stage answers,
without the marketing detour.
The questions every first call starts with — sovereignty, models, integrations, timelines, money. Sharper, longer answers live on /developers and /self-hosted; these are the homepage version.
- On infrastructure you choose. Self-hosted: inside your own data centre, your VPC, or an air-gapped rack — nothing leaves your perimeter. Arkintel-hosted: on European-headquartered, EU-resident providers we operate end-to-end, with your encryption keys. EU jurisdiction either way — no “we promise we won’t replicate to the US” footnote.