EU data residency
The European AI Gateway
Route Claude, GPT, Gemini and Mistral through EU-region endpoints. Hosted in Frankfurt. Zero retention. GDPR compliant. OpenAI SDK compatible.
One base URL and your Requesty key. No Azure, AWS or Vertex account needed.
Gateway regions eu-central-1
EU base URL
https://router.eu.requesty.ai/v1| Region | Serves | Policy |
|---|---|---|
| eu-central-1pinned | EU, Frankfurt | EU models only |
| us-east-1 | US, Virginia | refused |
| ap-southeast-1 | APAC, Singapore | refused |
The caller does not pick a region. The routing policy on the key decides, and a call that would leave it is refused.
Trusted by teams building in Europe
Compliance
Built for EU compliance from day one
Enterprise-grade AI gateway with full European data residency. No compromises on performance or security.
EU data residency
All data hosted exclusively in Frankfurt, Germany (AWS eu-central-1). No data ever leaves the European Union.
Zero data retention
Requests are proxied in real-time. Nothing is stored, logged, or cached after delivery. We never train on your data.
GDPR compliant
Full GDPR compliance with DPA on request. Complete audit logging available for enterprise customers.
Encryption and privacy
TLS 1.3 in transit, AES-256 at rest. PII detection and scrubbing. Prompt injection protection built-in.
OpenAI SDK compatible
One line to switch. Change your base_url to our EU endpoint and use your Requesty API key. Zero code changes.
All major providers
Anthropic, Azure OpenAI, Google Vertex AI, AWS Bedrock, Mistral, and Nebius. All routed through Frankfurt.
By model family
EU data residency, per model family
Every major model family has EU-region deployments. See which models, regions and prices are available for each.
Claude with EU data residency
Opus, Sonnet and Haiku on Bedrock and Vertex EU regions
OpenAI GPT with EU data residency
GPT-5 and GPT-4.1 on Azure France, Sweden and UK
Gemini with EU data residency
Gemini Flash on Vertex AI EU regions
Mistral with EU data residency
European models on EU-hosted infrastructure
Comparison
Requesty EU against going direct
What changes when the routing layer is in Frankfurt instead of in front of each provider separately.
EU compliance
| Capability | Requesty EU | Direct provider |
|---|---|---|
| EU-only data processing | YesFrankfurt | NoUS default |
| Zero data retention | Yes | NoLogs stored |
| GDPR compliant | YesDPA available | PartialVaries |
| PII scrubbing | YesBuilt-in | No |
| Prompt injection protection | Yes | No |
Gateway features
| Capability | Requesty EU | Direct provider |
|---|---|---|
| Multi-provider routing | Yes600+ models | No |
| Response caching | Yes60% savings | No |
| Automatic failover | Yes | No |
| Rate limit handling | Yes | Partial |
| Cost tracking and analytics | YesReal-time | Partial |
Setup and integration
| Capability | Requesty EU | Direct provider |
|---|---|---|
| Migration effort | Yes1 line change | NoSDK rewrite |
| OpenAI SDK compatible | Yes | Partial |
| Single EU endpoint | Yes | NoPer-provider setup |
| Free credits to start | Yes$10 | No |
| Custom invoicing (EUR) | Yes | Partial |
Integration
One line to EU compliance
OpenAI SDK compatible. Switch your AI infrastructure to EU in seconds.
- Change the base URL, keep the rest of your code.
- One Requesty key reaches every EU-region deployment.
- Switch model or leave with one line.
from openai import OpenAI
# Initialise against the EU gateway
client = OpenAI(
base_url="https://router.eu.requesty.ai/v1",
api_key="your-requesty-key",
)
# Every request is routed through Frankfurt
response = client.chat.completions.create(
model="anthropic/claude-opus-4-6",
messages=[{"role": "user", "content": "Hello!"}],
)The same call reaches Bedrock, Azure OpenAI, Vertex AI, Mistral and Nebius. Only base_url changes.
Questions
Frequently asked questions
Everything you need to know about EU data residency with Requesty.
Where is my data hosted and processed?
All data is hosted, stored, and processed exclusively in Frankfurt, Germany (AWS eu-central-1). No data ever leaves the European Union. Our entire infrastructure stack runs within EU borders.
Is Requesty DSGVO-konform? (GDPR-compliant in Germany)
Yes. Requesty is fully DSGVO/GDPR-compliant. The EU gateway runs entirely in Frankfurt with zero data retention. We offer a DPA (Auftragsverarbeitungsvertrag) covering Article 28 processor obligations. Unlike OpenRouter, which routes through US infrastructure, Requesty keeps all AI traffic within the EU.
What is the best GDPR-compliant AI gateway for European companies?
Requesty is a GDPR-compliant AI gateway hosted in Frankfurt, Germany. It routes requests to Claude, GPT, Gemini, and Mistral through EU-region endpoints with zero data retention. Features include automatic failover, cost optimization, PII scrubbing, and enterprise governance. One API key replaces separate accounts with each provider.
Is there a European alternative to OpenRouter?
Yes. Requesty is an EU-hosted AI gateway: the routing layer runs in Frankfurt, request and response bodies are never stored, and you can route to EU-region deployments of Claude, GPT, Gemini and Mistral. Both expose an OpenAI-compatible API, so migrating is a base-URL and API-key change.
Does Anthropic offer EU data residency for Claude?
Not through its first-party API. EU data residency for Claude is available through cloud deployments: AWS Bedrock EU regions (Frankfurt, Ireland, Paris, Stockholm) and Google Vertex AI EU regions, all routable through Requesty with one OpenAI-compatible API.
Can I run OpenAI GPT models with EU data residency?
Yes, via Azure OpenAI EU regions. GPT-5, GPT-5 Mini and the GPT-4.1 family run in France Central, Sweden Central and UK South, routable through the Requesty EU gateway without an Azure subscription.
What is the fastest AI endpoint with EU data residency?
The Requesty EU gateway adds roughly 8ms of routing latency in Frankfurt and connects to EU-region model deployments directly: Bedrock eu-central-1, Azure France Central and Sweden Central, Vertex EU regions. Model inference speed depends on the model you pick; the gateway overhead stays in single-digit milliseconds.
Do you store or retain any of my data?
No. Requesty operates with zero data retention. Requests are proxied to the model provider in real-time and nothing is stored, logged, or cached after delivery. We never train on your data.
Is Requesty GDPR compliant? Do you offer a DPA?
Yes. Requesty is fully GDPR compliant and offers a Data Processing Agreement covering Article 28 processor obligations, email sales@requesty.ai to request one. Complete audit logging is available, and our infrastructure is designed from the ground up for EU data residency.
Can I get custom invoicing for my company?
Absolutely. We support custom invoicing, PO numbers, and flexible billing cycles tailored to your procurement process. We can invoice in EUR and accommodate enterprise billing requirements. Contact our team to set this up.
Which AI models are available through the EU gateway?
We support all major model providers through EU-based endpoints including Anthropic, Azure OpenAI, Google Vertex AI, AWS Bedrock, Mistral, and Nebius. All traffic is routed exclusively through our Frankfurt infrastructure.
How do I migrate from OpenAI or another provider?
It takes one line of code. Simply change your base_url to our EU endpoint and use your Requesty API key. Our gateway is fully OpenAI SDK compatible.
What security features are included?
Every request passes through our security layer which includes PII detection and scrubbing, prompt injection protection, topic restriction policies, and custom regex patterns. These guardrails are configurable per team or project.
More on security and compliance
Provenance. Model count (600+) is a platform measurement. Uptime SLA of 99.99% is a contractual commitment, not an observed result. EU customer count, gateway latency and cache savings are carried over from the live /eu page and are not in the verified figure set; they need a source before this ships. Customer logos are the marks of Requesty customers as published on /customers. The wordmark here is live text rather than the outlined SVG production requires.
Your AI data stays in Europe
One base-URL change puts Claude, GPT, Gemini and Mistral behind an EU-pinned key.
DPA on request. Custom invoicing in EUR.