--- title: "EU AI inference platform · Claude, Gemini and open LLMs in Europe" description: "Run Claude, Gemini, open-source LLMs and AI agents on EU infrastructure, under one European contract and one DPA. Data stays in European data centres." canonical: "https://inference.eu/" language: "en" --- AI inference platform hosted in the EU # Run frontier and open AI models on EU infrastructure Claude, Gemini and open-weight LLMs behind one API, plus agents, integrations and GPU servers, all processed in European data centres. For companies that need AI without sending data to the US. [Contact sales](https://inference.eu/contact/)[See what's included](#platform) 1. 01One DPA for every model and agent 2. 02Claude, Gemini and open models 3. 03European contract, invoiced in euros ## Everything you need to run AI in the EU Start with an API call to an open model. Add agents, connections to your apps, and your own GPUs when you need them. All of it runs in the same European regions, under one contract. - [ ### Models Claude, Gemini and open-weight LLMs through one API, hosted in the EU. About Models](https://inference.eu/models/) - [ ### Agents AI agents that use your tools, run on a schedule and ask before they act. About Agents](https://inference.eu/agents/) - [ ### Connect Give AI agents access to your apps through MCP servers, OAuth and API keys. About Connect](https://inference.eu/connect/) - [ ### Skills Reusable instructions and scripts that teach AI agents a specific task. About Skills](https://inference.eu/skills/) - [ ### GPUs Dedicated GPU servers in European data centres, rented by the hour. About GPUs](https://inference.eu/gpus/) - [ ### Not sure where to start? Tell us what you're building and we'll suggest the shortest path, from a single API call to a full agent setup. Talk to us](https://inference.eu/contact/) ## Switch from OpenAI by changing one line The API uses the OpenAI request format, so the SDKs and frameworks you already use work as they are. Change the base URL and the key, then compare models on your own prompts. ``` from openai import OpenAI client = OpenAI( base_url="https://api.inference.eu/v1", api_key="INFERENCE_EU_API_KEY", ) response = client.chat.completions.create( model="mistral-small-4-119b", messages=[{"role": "user", "content": "Summarise this contract in three bullet points"}], ) print(response.choices[0].message.content) ``` ``` import OpenAI from "openai" const client = new OpenAI({ baseURL: "https://api.inference.eu/v1", apiKey: process.env.INFERENCE_EU_API_KEY, }) const response = await client.chat.completions.create({ model: "mistral-small-4-119b", messages: [{ role: "user", content: "Summarise this contract in three bullet points" }], }) console.log(response.choices[0].message.content) ``` ``` curl https://api.inference.eu/v1/chat/completions \ -H "Authorization: Bearer $INFERENCE_EU_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "mistral-small-4-119b", "messages": [{"role": "user", "content": "Summarise this contract in three bullet points"}] }' ``` ## Frontier and open models for every task Chat, code, reasoning, embeddings, reranking, speech and images, all behind one endpoint. [All models](https://inference.eu/models/#directory) - [Claude Sonnet 5.5Use Anthropic's Claude Sonnet 5.5 on inference.eu: requests processed in EU data centres, one data processing agreement, nothing used for training.Chat](https://inference.eu/models/claude-sonnet-5-5/) - [Gemini 4 ArgonUse Google's Gemini 4 Argon on inference.eu: requests processed in EU data centres, one data processing agreement, and nothing used for training.Reasoning](https://inference.eu/models/gemini-4-argon/) - [gpt-oss-120bRun OpenAI's open-weight gpt-oss-120b reasoning model on inference.eu: the OpenAI API format, with requests processed in EU data centres, billed in euros.Reasoning](https://inference.eu/models/gpt-oss-120b/) - [Llama 3.3 70B InstructRun Meta's Llama 3.3 70B Instruct on inference.eu: an OpenAI-compatible API with requests processed in EU data centres and billed in euros.Chat](https://inference.eu/models/llama-3-3-70b-instruct/) - [Qwen3 235B A22B InstructRun Qwen3 235B A22B Instruct (2507) on inference.eu: a 256K-context open model behind an OpenAI-compatible API, processed in EU data centres.Chat](https://inference.eu/models/qwen3-235b-a22b-instruct/) - [DeepSeek V4.1 Flash (1017)Run DeepSeek V4.1 Flash (1017) on inference.eu: a multimodal open reasoning model with 1M context, OpenAI-compatible, processed in the EU, not in China.Reasoning](https://inference.eu/models/deepseek-v4-1-flash-1017/) - [Kimi K3Run Kimi K3, Moonshot AI's 2.8-trillion-parameter open-weight model with 1M context and image and video input, on inference.eu with processing in the EU.Reasoning](https://inference.eu/models/kimi-k3/) - [GLM-5.3Run Z.ai's GLM-5.3, an open-weight model built for complex coding and long agent tasks, on inference.eu with OpenAI and Anthropic APIs, processed in the EU.Reasoning](https://inference.eu/models/glm-5-3/) ## Your data stays in the EU The questions your legal and security teams will ask, answered before they ask them. - ### Processed in European data centres Prompts and responses are handled in EU locations. Backup capacity is in the EU as well, so no request is ever routed to another region. - ### Never used for training What you send answers your request and nothing else. It is not used to train any model. - ### One data processing agreement A single DPA covers every model, agent and connection, with the list of subprocessors and locations. - ### Contracted in Europe Your contract, invoices in euros and support sit with a European entity. ## Works with the tools your developers use Step-by-step setup for coding agents, IDEs, SDKs and chat apps. [All integrations](https://inference.eu/integrations/) - [Claude CodePoint Claude Code at inference.eu with ANTHROPIC\_BASE\_URL and run it on open-weight models processed in EU data centres. No Anthropic account needed.Coding agents](https://inference.eu/integrations/claude-code/) - [OpenCodeRun the OpenCode terminal agent on open-weight models processed in EU data centres. Add one provider block to opencode.json and keep your code in Europe.Coding agents](https://inference.eu/integrations/opencode/) - [CursorUse open-weight models processed in EU data centres for Cursor Chat and Agent through the OpenAI base URL override. What it covers and what still goes via Cursor.IDEs and editors](https://inference.eu/integrations/cursor/) - [OpenAI Codex CLIRun OpenAI's Codex CLI on open-weight models processed in EU data centres. Add a model\_providers block to config.toml and keep code in Europe.Coding agents](https://inference.eu/integrations/openai-codex/) - [Vercel AI SDKUse the Vercel AI SDK's OpenAI-compatible provider to call open-weight models processed in EU data centres from Next.js, Node.js or edge apps.SDKs](https://inference.eu/integrations/vercel-ai-sdk/) - [LangChainConnect LangChain's ChatOpenAI to inference.eu with base\_url and run chains, agents and LangGraph workflows on open-weight models processed in the EU.Frameworks](https://inference.eu/integrations/langchain/) - [n8nPoint n8n's OpenAI credential at inference.eu and run AI Agent and Chat Model workflows on open-weight models processed in EU data centres.Automation](https://inference.eu/integrations/n8n/) - [Open WebUIConnect Open WebUI to inference.eu as an OpenAI-compatible connection and give your team a self-hosted chat on open-weight models processed in the EU.Chat apps](https://inference.eu/integrations/open-webui/) ## Built for teams that can't send data to the US If your customers, regulators or security team ask where your AI runs, this is for you. - ### SaaS teams adding AI features Ship chat, summaries and search inside your product without a US AI provider appearing in your customers' security reviews. - An EU subprocessor you can name in your DPA - The OpenAI SDK you already use - Per-token pricing in euros - ### Regulated companies Finance, health, legal and the public sector, where data residency is a requirement rather than a preference. - Processing kept in EU data centres - Requests never used for training - Approval steps before agents act - ### Teams leaving US AI providers Keep your code and prompts, point them at an EU endpoint, and compare quality model by model before you switch traffic. - One line of code to switch - Run both side by side while you test - Help choosing the closest open model ## MCP servers for European business apps Let agents work in DATEV, Personio, Pennylane, Qonto, Légifrance and the tools you use every day. [All MCP servers](https://inference.eu/connect/#directory) - [DATEVConnect inference.eu agents to DATEV through an MCP server. Read postings, clients and documents, and prepare uncommitted booking batches for review.Accounting](https://inference.eu/connect/datev/) - [PersonioConnect inference.eu agents to the official Personio MCP server. Read employee, absence and compensation data and request time off, within each user's rights.HR and payroll](https://inference.eu/connect/personio/) - [sevdeskConnect inference.eu agents to sevdesk through an open-source MCP server. Audit VAT, reconcile bank payments and draft invoices, read-only by default.Accounting](https://inference.eu/connect/sevdesk/) - [SAP S/4HANAConnect inference.eu agents to SAP S/4HANA or ECC through an MCP server that exposes OData services via SAP BTP. Query and update business data with approvals.ERP](https://inference.eu/connect/sap/) - [PennylaneConnect inference.eu agents to the official Pennylane MCP server. Read invoices, ledgers, trial balances and VAT returns for one company or a whole firm.Accounting](https://inference.eu/connect/pennylane/) - [GitHubConnect inference.eu agents to GitHub's official MCP server. Read code, triage issues, review pull requests and check Actions runs, with read-only mode available.Developer tools](https://inference.eu/connect/github/) - [SlackConnect inference.eu agents to Slack's official MCP server. Search messages and files, read channels and threads, post updates and edit canvases with admin approval.Communication](https://inference.eu/connect/slack/) - [NotionConnect inference.eu agents to Notion's official MCP server to search, read, create and update pages and databases in your workspace, with approvals.Productivity](https://inference.eu/connect/notion/) ## Up and running in a day Access is set up per company, so we can agree models, volume and the DPA with you first. 1. 1 ### Tell us what you're building Which models you need, roughly how much traffic, and any compliance requirements. 2 minutes 2. 2 ### We reply within one working day With answers to your questions, pricing for your volume, and a call if it helps. 1 working day 3. 3 ### You get an API key and a DPA Start calling models straight away, with a data processing agreement ready to sign. Same day ## Common questions The questions buyers, developers and legal teams ask us most. Something missing? Ask us directly. A platform for running frontier and open-source AI models, agents and the integrations they use on EU infrastructure, through one OpenAI-compatible API and one contract. No. If you use an OpenAI SDK or a framework built on one, such as LangChain, LlamaIndex or the Vercel AI SDK, change the base URL and the API key. Prompts, tools and streaming work as before. In data centres in the European Union, including the backup capacity we switch to during an outage. The full list of locations and subprocessors comes with your DPA. No. What you send is used only to answer your request, never to train a model. Frontier models such as Claude and Gemini, and open-weight models for chat, code, reasoning, embeddings, reranking, speech and images, including gpt-oss, Llama, Qwen, DeepSeek, Kimi, GLM, Mistral and Gemma. Browse the full list on the models page. Models are priced per token and billed in euros on one invoice. For steady or high volume we quote a monthly price for your usage. We provide a data processing agreement under Article 28 of the GDPR, process requests in the EU, and list every subprocessor. Whether your use is compliant also depends on how you use it, and we'll answer your legal team's questions. Your contract is with a European entity and requests are processed in European data centres. Ask us for the details of our ownership and subprocessors, and your legal team can assess the exposure themselves. Yes. Agents run on the same platform, can call your apps through OAuth, API keys or MCP servers, and can ask a person to approve anything risky before they act. Yes. The API speaks both the OpenAI and the Anthropic formats, so coding agents and IDEs that let you set a custom endpoint work with EU-hosted open models. Each integration page has the exact setup. Yes. For fine-tuning, serving your own weights or batch jobs you can rent dedicated GPU servers in EU regions by the hour. Send us your work email and what you're building. We reply within one working day with pricing for your volume and an API key. [Ask a question](https://inference.eu/contact/) ## Move your AI workloads to EU infrastructure Tell us which models you use and roughly how much traffic you have. We reply within one working day with pricing and next steps. - Reply within one working day - Pricing for your volume - DPA ready to sign