01/Overview

Getting Started

Send your first request in a few minutes.

The gateway is OpenAI-compatible: connect with your existing OpenAI client by only changing the base URL and the key.

#Authentication

Requests are authenticated with a Bearer token. When you log in as a user, the panel gives you an API token; send it in the Authorization header on every request.

Authorization

Authorization: Bearer <YOUR_TOKEN>

#Your first request

Put either a model name or a model group name in the model field. If you use a group name, the group's strategy (e.g. flow) decides which real model is used.

curl

curl https://<gateway-host>/v1/chat/completions \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "my-group",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

python (openai)

from openai import OpenAI

client = OpenAI(base_url="https://<gateway-host>/v1", api_key="$TOKEN")
resp = client.chat.completions.create(
    model="my-group",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Streaming, tools/function calling and vision work as-is if the model supports them. Ollama-native paths (/api/chat, /api/generate) are supported too.