Host your AI in 5 minutes with a drop-in API.
Grow to a dedicated model on an isolated GPU cluster.
Keep your tools. Choose your models. Control your data and spending. We run the infrastructure in Europe.
GLM
DeepSeek
Qwen
Kimi
Mistral
Llama
gpt-oss
Gemma
Loes
Nemotron
Claude Code
Codex
OpenCode
Cursor
Cline
Loes
OpenAI SDK
Anthropic SDKEurope's teams build on AI that someone else controls. We think a team should choose its tools, its models and where its data is stored. That is why we run AI infrastructure in Europe: AI on our own terms.
One platform, four layers
Start with one layer and add the others when the work asks for it. They share one key, one balance and one Activity view.
01Code
Your coding agent on open models in Europe, set up with one command.
$ curl -fsSL https://hostyourai.com/cli/install.sh | sh
$ hyai
hyai code: OpenCode on model glm-5.3
(best available from the preference list)
02Control
Keys, budgets and guardrails per team, and personal data stopped before any model sees it.
What the model sees
Refund order 4471 for Sanne Mulder, IBAN NL91 ABNA 0417 1643 00
Personal dataoffflagblock
03Serve
Shared per token, or running for you alone, from European data centres.
04EU compute
Private GPUs in Europe, H100 to B300, sized for response time, generation speed and concurrent users.
| Model | State | GPUs | Region |
|---|---|---|---|
|
GLM 5.3 | Running | 8× H200 | Helsinki |
|
Qwen3.8 27B | Running | 1× H100 | Paris |
|
Llama 3.3 70B | Running | 2× H100 | Frankfurt |
One key.
Your rules check every request.
Your coding agent on open models in Europe, set up with one command.
hyai sets the base URL and key, then starts your agent. Cursor takes the base URL.
The route
Your engineers keep the tools they already use and connect them with one key. HostYourAI checks every request against your team's budget and rules, picks the open model that fits the work, and runs it on GPUs in Europe.
The route: your tools, one key, HostYourAI with your team's rules, three kinds of model, GPUs in Europe
Start alone on Serve. Add Control when a team joins. Move to EU compute when AI becomes core work. Same key, same tools, nothing to rebuild.
Serve
For one engineer or a side project.
Pay per token from prepaid credit
Get your API keyControl
For teams that want overview and control.
Per seat plus tokens, one bill split per team
Start a teamEU compute
For companies where AI is core work.
Per GPU minute, or a fixed monthly price
Talk to usPublished price per model on the pricing page. Every account gets a DPA.
Today we finetune the best open models with Dutch and European data and run them in Europe. Alongside, we build our own, so that one day Europe runs on models of its own.
We test every new open model on real work: code, analysis, calculation and writing. Our own suite plus public benchmarks, measured without our layer. The best model for each kind of work is the one your team gets.
We finetune the strongest open models with Dutch and European data, in Europe, with open European teachers. There is always a fully European track next to the strongest open model. Together they are Loes.
Alongside, we train our own base model, Baby Loes, from the first token in Europe. When it is good enough, it joins the benchmark like every other model.
Start with one key today. Add dedicated GPUs when your workload needs them.
$ curl -fsSL https://hostyourai.com/cli/install.sh | sh
Run from night to day through Code, Control, Serve and EU compute, and grab tokens and keys on the way.
Press space or tap to playSpace or up to jump, twice to go higher.
Your best score stays on this device.
Research clusterservice key
Reports agentFinance agent
Stanstan laptop
Helpdesk botSupport agent
Daandaan laptop
CI pipelinePlatform agent
Sandersander laptop
Kyrakyra laptop
Checkout appMobile agent
Data agentData science agent
Hans
Jantine
