A European AI gateway for frontier and open models
Standard Compute gives European companies Claude, GPT, Gemini and leading open models through one API. With a Europe rule, your prompts are processed only in Europe. They enter and leave through our gateway in Finland and are served only by approved European inference endpoints, with no fallback outside Europe.
Create a company workspaceor write to contact@standardcompute.com
Short answers
- Where are our prompts processed?
- Set a Europe rule for your company, a team or a single key, and every request, including retries and fallbacks, is served only by approved European inference endpoints. If no European endpoint can serve it, the request is refused before your prompt is sent.
- Where does the gateway run and where is our account data?
- Our EU gateway terminates TLS and runs in Finland, and your account data is stored in Frankfurt. Request records hold metadata only (model, provider, tokens, cost and status), never your prompts or outputs.
- Which frontier models are served in Europe?
- Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 on Google Vertex AI in Europe and Amazon Bedrock in Ireland, GPT-6.1 and GPT-5.6 in the Microsoft Azure EU Data Zone, and Gemini 3.5 Flash-Lite on Google Vertex AI. GLM, DeepSeek and MiniMax are also served by European providers.
- Do you keep or train on our data?
- We process your prompts in memory to serve each request and never write them to storage or logs. Standard Compute never trains on your prompts or outputs, and with a region rule or zero retention turned on, every request goes only to providers that commit not to train on it.
- What does our DPO get?
- Sign our Data Processing Agreement in the dashboard in minutes. It commits us to no training on your data, no retention of your content, and breach notification within 72 hours. We publish every sub-processor with its purpose, data and location, and give you 14 days' notice before adding a new one.
How processing in Europe works
Point your applications at our EU gateway, https://eu.stdcmpt.com/v1, and set a Europe rule on the company, a team or a key in the dashboard. Teams and keys inherit the rule, and a lower level can only narrow it.
Every prompt-bearing call is pinned to reviewed European endpoints: the routing step, the answer, every retry and every fallback. A model with no European endpoint is refused before any content is sent.
Region rules and the zero-retention requirement are free on every company workspace.
curl https://eu.stdcmpt.com/v1/chat/completions \
-H "Authorization: Bearer $STANDARD_COMPUTE_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "anthropic/claude-sonnet-5.5",
"messages": [{"role": "user", "content": "Summarise this contract clause."}]
}'Models served in Europe
| Model | European inference providers |
|---|---|
| Claude Sonnet 5.5 | European inference providers: Google Vertex AI |
| Claude Sonnet 5 | European inference providers: Google Vertex AI, Amazon Bedrock |
| Claude Opus 5.5 | European inference providers: Google Vertex AI, Amazon Bedrock |
| Claude Opus 5 | European inference providers: Google Vertex AI, Amazon Bedrock |
| Claude Haiku 5.5 | European inference providers: Google Vertex AI, Amazon Bedrock |
| GPT-6.1 Terra | European inference providers: Microsoft Azure |
| GPT-6.1 Sol | European inference providers: Microsoft Azure |
| GPT-5.6 Terra | European inference providers: Microsoft Azure |
| GPT-5.6 Sol | European inference providers: Microsoft Azure |
| GPT-5.6 Luna | European inference providers: Microsoft Azure |
| Gemini 3.5 Flash-Lite | European inference providers: Google Vertex AI |
| GLM 5.3 Flash | European inference providers: Inceptron, TensorX |
| GLM 5.3 | European inference providers: Mistral, Inceptron, TensorX |
| GLM 5.2 | European inference providers: Mistral, Inceptron, TensorX |
| DeepSeek V4.1 Flash 1017 | European inference providers: TensorX |
| DeepSeek V4.1 Flash | European inference providers: TensorX |
| DeepSeek V4 Pro 0813 | European inference providers: TensorX |
| DeepSeek V4 Flash 0731 | European inference providers: Inceptron, TensorX |
| MiniMax M3.1 Flash | European inference providers: TensorX |
| MiniMax M3 | European inference providers: TensorX |
From the endpoint list the gateway enforces, reviewed 2026-10-27. Google Vertex AI serves from its Europe locations, Amazon Bedrock from eu-west-1 in Ireland, and Microsoft Azure from its EU Data Zone.
Retention, training and records
Turn on zero data retention for your company, a team or a key, and every request is routed only to inference endpoints that commit to keeping no copy of your data. Teams and keys inherit the requirement and can never switch it off underneath you.
Choose exactly which inference providers may handle your data, for the whole company, a team or a single key. New providers are never added without an admin's approval, and fallbacks can never route around your list.
Every admin change is recorded with before and after values, and you can export a record of every request showing the model, the provider that served it and the region rule in force, kept for 12 months.
What a security review gets
Our security overview covers what we store, where it is stored, encryption and access control. The sub-processor list names every sub-processor with its purpose, data and location. The DPA is public, and you sign it in the dashboard.
Connect Microsoft Entra ID or any SAML identity provider so your people sign in to Standard Compute with their work accounts. Employees sign in to Claude Desktop and Claude Code with their Microsoft work account, through our EU gateway, so there are no API keys to hand out.
When each European option fits
European companies usually compare a gateway like ours with a European model vendor, a European cloud's model service, or a hyperscaler's EU region. Each fits a different need. The facts on the other providers come from their own pages.
| Option | Where inference runs, as stated | Models | Fits when |
|---|---|---|---|
| Standard Compute | Where inference runs, as stated: Approved European endpoints under a Europe rule; gateway in Finland | Models: Claude, GPT, Gemini, GLM, DeepSeek, MiniMax | Fits when: You want frontier and open models under one account, with rules per team and a budget for every key |
| Mistral AI | Where inference runs, as stated: Data hosted in the EU by default | Models: Mistral's own models and some open models | Fits when: Mistral's models cover your work and you want a vendor based in Paris |
| Scaleway Generative APIs | Where inference runs, as stated: Paris, France | Models: Open models such as GLM, DeepSeek, Qwen and Llama | Fits when: You want open models served in France by a French cloud |
| IONOS AI Model Hub | Where inference runs, as stated: Data centers in Germany | Models: Open-source models | Fits when: You want open models in German data centers |
| OVHcloud AI Endpoints | Where inference runs, as stated: Hosted and run in the EU | Models: 20+ open-weight models | Fits when: You already run on OVHcloud and need open models |
| Amazon Bedrock, EU inference profile | Where inference runs, as stated: EU Regions | Models: Claude and other models on AWS | Fits when: You are on AWS and manage quotas and accounts there |
| Microsoft Foundry, EU Data Zone | Where inference runs, as stated: EU Data Boundary | Models: GPT models; Claude's Data Zone option is US only | Fits when: You are a Microsoft shop and GPT covers your work |
Checked on each provider's own pages on 2026-11-07. Sources are listed at the end of this page.
Questions from IT, security and procurement
- Which LLM API providers guarantee inference stays in Europe, not just storage?
- Standard Compute does. Under a Europe rule the inference itself runs in Europe: the routing step, the answer and every fallback run at approved European endpoints.
- Is there a GDPR-friendly LLM API with both Claude and GPT processed in Europe, with a DPA and a sub-processor list?
- Yes. Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 and GPT-6.1 are served by approved European endpoints under a Europe rule. Our DPA is signed in the dashboard and the sub-processor list is public at /sub-processors.
- Does "Europe" mean the EU only?
- Approved European endpoints are in the EU/EEA, the United Kingdom or Switzerland. Our EU gateway runs in Finland and your account data is stored in Frankfurt.
- Can we use Claude in Europe for health or legal data?
- Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 are served in Europe by Google Vertex AI and Amazon Bedrock in Ireland. Add zero data retention, sign the DPA, and request records hold metadata only, never prompts or outputs.
- What happens when a model has no European endpoint?
- The request is refused before your prompt is sent. A Europe rule never falls back to an endpoint outside Europe.
- Is this sovereign AI? What about the US CLOUD Act?
- You choose where inference runs and which providers may serve you. Frontier models in Europe are served from the European regions of Google, AWS and Microsoft; if your policy accepts only European-headquartered providers, an allow-list limits a key to them, for example Mistral for GLM models. Our DPA is governed by Norwegian law.
- Do we need a separate contract with every model vendor?
- No. One company account gives you every model, and you sign one DPA with us. The inference providers that serve Europe-restricted workspaces are listed on /sub-processors.
- Can we set this up for one department first?
- Yes. Organize your company into teams, invite colleagues by email, and give every person and every app its own key. Set the Europe rule on one team and widen it to the company when you are ready.
- What are the European alternatives to the OpenAI and Anthropic APIs for companies that want sovereignty? Mistral, Aleph Alpha, who else?
- Standard Compute is a European AI gateway from a Norwegian company: one API for Claude, GPT, Gemini and open models, with a Europe rule that serves every request only from approved European inference endpoints and refuses it before your prompt is sent if none qualifies. If your policy accepts only European-headquartered providers, a provider allow-list limits a key to them, for example Mistral for GLM models.
Related pages
- Standard Compute for companies
- Security overview
- Open-source models for business
- Sovereign AI: self-hosting or a European gateway
- EU LLM API for coding agents
- LLM-API nach DSGVO (Deutsch)
- API LLM et RGPD (Français)
- LLM-API og GDPR i Norge (Norsk)
- Sub-processors
- Data Processing Agreement
Sources
- Mistral AI: where your data is stored, checked 2026-11-07
- Mistral AI: models, checked 2026-11-07
- Scaleway Generative APIs: FAQ, checked 2026-11-07
- IONOS AI Model Hub: data handling, checked 2026-11-07
- OVHcloud AI Endpoints, checked 2026-11-07
- Amazon Bedrock: geographic cross-Region inference, checked 2026-11-07
- Microsoft Foundry: deployment types (Data Zone), checked 2026-11-07
- Microsoft Foundry: Claude hosting options, checked 2026-11-07
Create your company, set the Europe rule and invite your first team from the dashboard.