Start building
The base URL, your key, and a first call in curl and Python. About a minute end to end.
/docs/quickstart/AxForge Dev
Build on European AI infrastructure
An OpenAI-compatible API, hosted in the EU with zero prompt retention. Point the SDK you already use at our base URL, swap the key and the model name, and chat, vision, embeddings, images and speech all work behind one key.
eu-se-1 · Stockholm OpenAI-compatible Zero prompt retention
Start here
Connect your tool
Setups for the OpenAI SDK, the Vercel AI SDK, Cline, Continue, LiteLLM, aider, Open WebUI and more.
/docs/connect/API reference
Chat with streaming, tool calls, reasoning and image input, plus embeddings, images and speech.
/docs/chat/Run the chatbox starter
A tiny streaming chatbox in plain JavaScript. Paste your key and it runs — no framework, no build step.
/docs/starters/Test a call
Run a real call and see how it lands: connected, key and quota, which model, how long — and if it fails, why.
/docs/test-a-call/When something fails
Every status code, what causes it, and what to do about it.
/docs/errors-limits/Everything behind one key
Chat, agents & tools
Streaming, tool calls, reasoning and structured JSON output on /v1/chat/completions.
Vision & documents
Send images and scans — up to eight per request — and pull fields out of invoices and forms as JSON.
Embeddings & RAG
Vectors for search and retrieval on /v1/embeddings, same key and region as chat.
Images
Generate PNGs and make instruction edits to images you already have.
Speech & music
Speech to text, text to speech and music generation.
Models & pricing
The models we serve, what each is for, and how per-token pricing works.
Why build here
Hosted in the EU
Inference runs in-region on hardware we operate. A named region, pinned per key.
Zero prompt retention
Prompts and completions are not written to disk, logged, or trained on.
Change two lines
Point your existing OpenAI client at our base URL and swap the key. Most SDKs need nothing else.