Getting Started with the OpenAI-Compatible API
Because thousands of tools already speak the OpenAI protocol, you can plug them in without writing any special integration code. Same JSON structure, same authentication header, same streaming events.
Example with curl 🖥️
The fastest way to test your token is a single curl request:
curl https://api.your-domain.com/v1/chat/completions -H "Authorization: Bearer sk-your-token" -H "Content-Type: application/json" -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello"}]}'
Example with Python 🐍
The official openai Python library works out of the box. Just point base_url to our endpoint and set your API key.
Supported frameworks table
| Framework | base_url option | Works out of the box |
|---|---|---|
| openai-python | openai.base_url | Yes ✅ |
| openai-node | baseURL | Yes ✅ |
| LangChain | openai_api_base | Yes ✅ |
| LiteLLM | api_base | Yes ✅ |
| Open WebUI | OPENAI_API_BASE_URL | Yes ✅ |
What about LangChain, and other frameworks? 🧩
All major frameworks let you configure a custom base URL and API key. Look for the base_url option and set it to our endpoint. It really is that simple.
Frequently asked questions
Is the endpoint truly OpenAI-compatible?
Yes. Requests, responses and streaming events match the OpenAI protocol, so tools work without modification.
Which SDKs can I use?
Any library that lets you set a base URL: openai-python, openai-node, LangChain, LiteLLM, Open WebUI and more.
Do I need a special API key format?
No. A normal Bearer token created in the panel works with every SDK.
What happens if I send a wrong model name?
You get a model_not_found error that lists the models available on your account.