AI can build you a demo. We build you a product.

From a first version to a launch-ready MVP: web apps, mobile apps, APIs and dashboards. For a hardware product, we also build the app and the website that go with the device.

From one week. An MVP is a launch-ready first version of your web or mobile app: design, build, automated tests, deployment and handover. The MVP package costs from €299.99 excl. VAT, starts with Day One, and its final price is fixed in writing before any work starts.

Yes. We write code and documentation another developer can pick up, and we hand everything over at the end of a project.

In the EU. What we host for you runs in the EU data centres of Scaleway or OVHcloud, European cloud providers, and your data stays within the EU-27 unless you ask otherwise. For public-sector and critical-service projects, we design to the rules you answer to, such as GDPR and NIS2, and document how the system meets them.

You can take everything and run it yourself, or choose a Care plan: fixes, updates, monitoring and EU hosting from the engineers who built it.

Engineers write the plan and AI drafts code against it. A senior engineer reviews every line AI writes, and the code is tested and checked for security before it ships. The AI models we use don't train on your code or data.

Yes, it's one of our offers. Day One starts with a production-readiness review: what's solid, what's risky and what it takes to go live. We fix the most critical issues the same day.

Yes: language-model features, search over your own data, and automation, built with cost and privacy in mind. For data that has to stay in-house, see Local AI.

Yes. We install open-weight models such as Qwen, DeepSeek, GLM, Kimi, Llama and Gemma on hardware in your premises, sized to the models you need and the people who use them. Models run fully offline, so no prompt or document leaves your network. Installation starts from €999.99 excl. VAT, plus the hardware.

If you'd rather not run the hardware yourself, we can run the same models for you on an EU sovereign cloud, in the EU data centres of Scaleway or OVHcloud, from €999.99/month excl. VAT, with no set-up fee. The monthly price includes dedicated GPUs, model set-up, monitoring and support, with updates to newer models when the hardware supports them, and there are no usage limits beyond what the hardware can handle. The cloud is billed monthly, with a minimum commitment of 6 to 12 months, depending on the setup. Either way, nothing is sent to the model's publisher.

It depends on the models you need and the people who use them. We offer three setups:

• Compact: one workstation with a single GPU (24–⁠48 GB of GPU memory), or a Mac Studio with at least 64 GB of memory. Models: Gemma, smaller Qwen models, Llama 8B.
• Department: one server with 2 to 4 data-centre GPUs (192–⁠384 GB of GPU memory). Models: smaller GLM models, larger Qwen models, Llama 70B.
• Organisation: one or more 8-GPU servers, more than 1 TB of GPU memory each; the largest models need two or more. Models: DeepSeek, Kimi, GLM, the largest open-⁠weight Qwen models.

We check each model's licence against your use before we install it. Each setup runs on your premises or on an EU sovereign cloud: see Local AI setups for who each one suits and what it costs.

Yes: electronics, firmware, apps and websites, with prototypes assembled in our own workshop in Luxembourg. A first prototype usually takes weeks. A market-ready product takes months, sometimes longer: the smart speaker we built for a client took 1.5 years from start to delivery.

The device, its firmware, the app and the website, from one team. We design boards, source parts and assemble prototypes in our own workshop. We build on ESP32, STM32 and Arduino, with circuit boards fabricated by JLCPCB and PCBWay and parts from Mouser and Digikey.

In the workshop, a pick-and-place machine, a reflow oven and 3D printers let us assemble, solder and fit small batches of prototypes into printed enclosures in-house.

With Day One: the architecture, a component shortlist with a cost estimate, a proof of concept on a development board where feasible, plus the implementation plan. The Hardware product package costs from €1,499.99 excl. VAT, and Day One is credited toward it if you continue within two weeks of paying the Day One invoice.

Yes. We can join your engineers or run a project end to end. Either way, a senior engineer reviews every line AI writes before it ships. We work with you as consultants, at a day rate, starting with Day One.

No. We also join engineering teams and research labs, and our EU hosting and Local AI are made for companies and public bodies that can't send their data just anywhere.

Both. Our base and workshop are in Frisange, just south of Luxembourg City. We come on site around Luxembourg and across the EU, the UK, Norway and Switzerland; in North America, the Middle East and North Africa, we work with you remotely day to day and travel to you when your project needs it.

See where we work on the map.

Day One costs €199.99 excl. VAT. Prices for everything else are in the pricing section. After Day One you get a fixed price in writing for the next step, and nothing starts without your OK.

There's no payment on this site. If it helps, we start with a meeting to talk your project through. Then we confirm Day One and send its invoice by email, and you pay it before the day.

For every step after it, the price is fixed in writing before any work starts, and nothing is booked or charged until you say yes. The MVP, Hardware product and AI-built app to production packages are paid by milestone. Care is billed monthly. The EU sovereign cloud is billed monthly, with a minimum commitment of 6 to 12 months, depending on the setup.

You pay by bank transfer or credit card. Prices exclude VAT.

Yes, if you continue within two weeks of paying the Day One invoice: what you paid for Day One is credited toward whatever you build next with us. If you stop there, you keep everything we produced.