Independent company · we are not affiliated with, sponsored by, or endorsed by the Swiss AI Initiative (EPFL, ETH Zurich, CSCS), the developers of the Apertus model · Official site: apertus-ai.org
Dedicated AI & MCP gateway for enterprises
apertus.ai runs your AI & MCP gateway dedicated to your company, in ISO 27001 certified data centres in the EU or on your premises. Every employee, tool and agent gets secure access to vetted models and your data, without a shared platform.
01–03 What we do
From your own gateway to open-source and custom apps to operations in the EU or in your building: all dedicated to your company, all from one team.
Not just another API key on a shared platform: we deploy your AI & MCP gateway per customer, with European models out of the box and controlled access to your tools.
We host and maintain proven open-source solutions such as Open WebUI, n8n and Qdrant, and deploy your own applications on the same Kubernetes platform.
We run everything in ISO 27001 certified data centres in the EU or on your premises, including local AI models on your own hardware.
04 Models where you need them
Through your gateway you use models in three modes, with the same API, the same permissions and the same audit log. You decide per use case where a model runs.
Cloud models
Curated open-weight models in ISO 27001 data centres in the EU, billed per token.
Models & pricingDedicated models
Reserved GPU capacity in the EU: guaranteed throughput, your own or fine-tuned models, no shared resources.
Compare the modesOn-premise
Local AI hardware in your building, delivered ready to run and maintained by us.
AI Box from €5,900
Explore local AIArchitecture How it fits together
Every request passes your own gateway. You see who uses AI, what it costs and which data goes where, and you switch models without touching applications.
Requests to language models
Tool calls from agents
Who uses AI
Your gateway · dedicated
LiveOne endpoint.
Every model.
One access point.
Every tool.
https://<your-gateway>.cloud.apertus.ai/v1
tools/call → checked · logged
European & local models
Your tools & data
Nº 03Underneath: managed operations by apertus.ai.
EU cloud · your cloud · on-premise
05 On-premise
When apertus.ai runs on your premises, the MCP gateway connects local models directly to your internal databases, ERP, DMS and file shares. Every access goes through permissions and the audit log, and no request ever leaves your network.
Your network
Your data sources
apertus.ai gateway
AI & MCP, on-premise
06 Integrations
Chat, coding agents or automation: everyone in your company reaches approved models and tools through your gateway, with their own keys, budgets and permissions. Anything that speaks an OpenAI- or Anthropic-compatible API or MCP works.
Principles Why apertus.ai
Deployment
1:1
Every customer gets their own gateway instance with its own endpoint. No shared platform, no other people's rate limits.
Location
EU
Operated in our European cloud or in your data centre, GDPR-compliant and without US dependency.
Endpoint
1
Models, tools and apps behind one access point, with permissions, cost and audit log in one place.
Lock-in
0
OpenAI-compatible API, open-weight models and open-source apps. You stay independent, including of us.
Rollout How we start
No big-bang project. We start with one concrete case and roll out once it works.
Together we look at which models, tools and applications you already use and where things get stuck today.
Result: a clear picture of where you stand
We deploy your gateway in the EU cloud or on your premises, with Open WebUI and the first use case. From then on usage, cost and access are visible.
Result: your gateway live
More teams, applications and agents join. Permissions, budgets and model choice stay with you, we run the operations.
Result: AI scales under control
FAQ Frequently asked
No. Every gateway is deployed per customer: own instance, own endpoint, own configuration. You are not just another API key on a shared platform.
Yes. We deliver your gateway fully preconfigured: models connected, teams, permissions and budgets set up, with Open WebUI if you like. The legal documents you need, such as the data processing agreement (DPA), come with it, so your privacy and procurement teams can sign off quickly.
No. The gateway is OpenAI-compatible. For your applications, usually only the endpoint changes, not the code.
Yes. Gateway, apps and local AI models can run entirely in your data centre, on your hardware or on our AI Box and AI Powerhouse. There, the MCP gateway connects the models securely to your internal databases.
European and open models for chat, coding, vision, embeddings and reranking are included out of the box, for example from Mistral, Swiss AI, DeepSeek, Qwen and Kimi. On-premise we run local open-weight models, closed-source models in the EU region are available on request.
In ISO 27001 certified data centres in the EU or on your premises, GDPR-compliant. You define with rules in the gateway which data may reach which model.
We price dedicated deployments of gateway, apps and operations per customer, please talk to our sales team. Models in our EU cloud are billed by token usage.
With anything that speaks an OpenAI- or Anthropic-compatible API or MCP: Open WebUI, Claude Code, Codex CLI, Cline, Continue, n8n, LangChain and many more. Every person gets their own keys and budgets.
Contact 20 minutes, an honest assessment
Book a short call. We look at your situation and tell you honestly whether and how apertus.ai can help.
What happens in the call
