The European alternative to the OpenAI API

Why do teams in the EU leave the OpenAI API?

Because an EU region doesn't make OpenAI European. OpenAI is a US company: its for-profit arm is a public benefit corporation controlled by OpenAI, Inc., a Delaware nonprofit corporation (Delaware Attorney General, 28 October 2025). US law follows US companies, not servers: the CLOUD Act obliges a US provider to hand over data in its possession, custody or control, wherever that data is stored (18 U.S.C. § 2713). For a European company that sends contracts, source code or customer data to an AI API, that is a strategic dependency, not only a compliance question.

OpenAI's own data controls page, checked on 28 October 2026, shows where the gaps are:

• EU data residency is by approval. It is a project setting for eligible customers, and OpenAI tells customers to ask its sales team whether they qualify. Any region outside the US also needs approval for abuse monitoring controls and a signed Modified Retention amendment.
• Part of your data still leaves the region. Data residency doesn't cover system data: account data, usage statistics, billing information, support requests and structured output schemas may be processed and stored outside the selected region.
• Prompts sit in abuse logs for 30 days. By default, OpenAI keeps abuse monitoring logs, which may contain prompts and responses, for up to 30 days. Zero Data Retention needs OpenAI's prior approval.
• Not every model or feature is covered. EU residency applies to the endpoints, models and snapshots OpenAI lists as eligible; Fast mode, for example, is not available with EU data residency for GPT-6 Astra or GPT-6 Sol.

To be fair to OpenAI: its API data has not been used to train OpenAI models since 1 March 2023, unless a customer opts in. The question for an EU team is not training. It is who controls the data, and under which law.

EUrouter answers each of these the other way around. EUrouter B.V. is an Amsterdam company, and its terms and DPA are governed by Dutch law. Every plan routes only to EU-hosted providers, with no sales call and no amendment, and personal data in API traffic is processed only in the EEA (EUrouter DPA, section 5).

What do you get with EUrouter?

EUrouter is an OpenAI-compatible API that is European by default: a European company, European infrastructure and mostly European providers, on every plan.

EU-hosted by default, on every plan

EUrouter sends inference traffic only to providers with EU-hosted endpoints, from the first request on the Free plan (routing docs). There is no residency setting to request and no plan to upgrade to. When a provider is down, EUrouter falls back to another EU-hosted provider of the same model, so a request never leaves Europe to stay online.

European companies all the way down

On 28 October 2026, 15 of the 17 providers in the EUrouter catalog are headquartered in the EU, among them Mistral AI, Scaleway, OVHcloud, IONOS and Nebul (providers). AWS Bedrock and Microsoft Foundry serve models from EU regions. When your policy asks for European companies only, set eu_owned: true on a request, or once in a routing rule set as your API keys' default (routing rules), and those two are excluded.

OpenAI models in Europe, and European models next to them

EUrouter serves 156 models from 26 labs on 28 October 2026, among them 23 OpenAI models and 23 Mistral models. GPT models run on Microsoft Foundry in EU regions. OpenAI's open-weight gpt-oss-120b runs on 9 providers, and 8 of them are European companies: AKI.IO, GreenPT, Infercom, IONOS, Nebul, OVHcloud, Regolo and Scaleway. The tables below read the live catalog, so models, prices and providers are always current.

OpenAI models, served from EU regions:

Mistral models on European providers:

Every model, with its providers and prices, is in the EUrouter model catalog.

No prompts stored

EUrouter doesn't store the content of requests or responses after a request completes; it keeps only metadata such as timestamps, model identifiers and token counts for billing (EUrouter DPA, section 10). All 17 providers are listed in the DPA with no prompt retention and no training.

How do you switch from the OpenAI API to EUrouter?

Switching from the OpenAI API to EUrouter takes two changes in an OpenAI SDK client: the base URL becomes https://api.eurouter.ai/api/v1 and the API key becomes an EUrouter key. Messages, streaming and tool calls stay the same (OpenAI SDK guide).

1. Create a free account at EUrouter sign-up and create a key under API Keys in the dashboard. EUrouter keys start with eur_ (quickstart).
2. Replace the base URL and the key in your client, as in the examples below.
3. Pick the model: keep a GPT model, or move to gpt-oss-120b or a Mistral model on European providers. Model IDs are in the model catalog.
4. Add the European controls your policy asks for: eu_owned, data_residency and max_retention_days (routing docs).

Python, with OpenAI's open-weight model on European providers only:

import os

from openai import OpenAI

client = OpenAI(
    base_url="https://api.eurouter.ai/api/v1",
    api_key=os.environ["EUROUTER_API_KEY"],
)

response = client.chat.completions.create(
    model="gpt-oss-120b",
    messages=[{"role": "user", "content": "Name three EU member states."}],
    extra_body={"provider": {"eu_owned": True}},
)

print(response.choices[0].message.content)

TypeScript, with a Mistral model:

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.eurouter.ai/api/v1",
  apiKey: process.env.EUROUTER_API_KEY,
});

const response = await client.chat.completions.create({
  model: "mistral-large-3",
  messages: [{ role: "user", content: "Name three EU member states." }],
});

console.log(response.choices[0].message.content);

The same request without an SDK:

curl https://api.eurouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $EUROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-oss-120b",
    "messages": [{"role": "user", "content": "Name three EU member states."}],
    "provider": {"eu_owned": true}
  }'

What changes and what stays, checked against the EUrouter API reference on 28 October 2026:

What other European options are there?

Besides EUrouter, three other routes keep your AI in Europe:

• Mistral AI (Paris, France): Mistral's own API, which hosts data in the European Union by default unless you choose its US endpoint (Mistral help center). A fit when Mistral's models are all you need.
• OVHcloud AI Endpoints (France): an OpenAI-compatible API for open models, hosted and run in the European Union (OVHcloud). A fit for teams already on OVHcloud.
• IONOS AI Model Hub (Germany): open-source models through an OpenAI-compatible REST API in IONOS Cloud data centers (IONOS). A fit for teams already on IONOS.

EUrouter routes to Mistral AI, OVHcloud and IONOS too, so you can use all three through one key and one contract. If your workload has no European requirement and you want OpenAI's newest models and features first, OpenAI's own API is built for that.

Moving away from OpenRouter instead? See the European alternative to OpenRouter, and for the difference between an EU region and European control, read data sovereignty for AI.

Start in Europe today

Create a free account, point your OpenAI client at https://api.eurouter.ai/api/v1, and your next request runs in Europe. Start free, read the routing docs, or talk to us about your DPA review.