On this page
Anthropic-compatible, EU-hosted
Want to keep using the Anthropic API but keep your inference in Europe? HostYourAI speaks the Anthropic API through a /v1/messages shim. You point your existing client at our Router and run open models on European GPUs.
What you get
- The
/v1/messagesshape withx-api-keyauthentication - Open models on EU infrastructure you control
- Also OpenAI-compatible through the same Router
- No training on your data; subprocessor list public
One platform for both APIs
If you run applications on both the OpenAI and the Anthropic API, you can do so through the same Router and the same EU GPUs. That keeps your stack simple and your data inside Europe.
How to switch
Migrating does not take weeks. Your existing Anthropic SDK keeps working: set the base URL to our Router and use your HostYourAI key as x-api-key.
from anthropic import Anthropic
client = Anthropic(
base_url="https://hostyourai.com",
api_key="hyai-...",
)
msg = client.messages.create(
model="llama-3.3-70b",
max_tokens=512,
messages=[{"role": "user", "content": "Hello"}],
)
print(msg.content[0].text)Pick an open model from the Model Garden, test in the Playground and track usage and cost per request. Want more detail? Read the guide Migrate your OpenAI client to the EU Router.
Claude API next to HostYourAI
| Claude API | HostYourAI | |
|---|---|---|
| Models | Claude, closed | Open: Llama, Qwen, Mistral, DeepSeek |
| API shape | /v1/messages | Same /v1/messages shape, plus the OpenAI route |
| Processing | United States | GPUs in the EU |
| Streaming | Token by token | Buffered via the Anthropic route; token by token via the OpenAI route |
| Images | Yes | Not yet via the Anthropic route, text-only |
| Auditability | Black box | Open weights |
Questions about the Claude API alternative
Do I get the same quality as Claude?
Closed frontier models lead on some tasks. For many business workloads, such as summarising, RAG and structured output, open models perform strongly and are auditable. Test your own use case honestly in the Playground.
Does my existing Anthropic code work?
Yes. You change the base URL and use your hyai- key as x-api-key. The /v1/messages shape stays the same; details are on the Anthropic-compatible API page.
What are the shim limitations?
Streaming is answered buffered and messages are text-only. If you need token-by-token streaming or multimodal input, use the OpenAI route on the same Router.
Why would I move away from Claude?
Usually data location: Anthropic processes in the US. With HostYourAI, prompts and outputs stay in the EU, with a Data Processing Agreement and no training on your data.
Can I migrate gradually?
Yes. Run both side by side and route per workload; the API shape is the same, so switching is a configuration change.
What does it cost?
Per token, per model class, from one prepaid balance. See the pricing page and the model catalog.
