Getting Started

New API is a unified AI model gateway. It exposes an OpenAI-compatible API and routes each request to one of your configured upstream channels.

Quick start

1. Deploy New API and open the web console.
2. Add a channel — point it at an upstream provider and paste its API key.
3. Create a token — this is the key your own clients will use.
4. Send a request to the gateway instead of the provider directly.

Your first request

Once a token exists, any OpenAI-compatible client works. Only base_url changes — it points at your gateway rather than at OpenAI:

curl http://localhost:3000/v1/chat/completions \
  -H "Authorization: Bearer sk-your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Next steps

• Channels — group upstreams and control routing by weight or priority.
• Tokens — scope keys to specific models and set quota limits.
• Logs — inspect per-request usage and troubleshoot failures.

Language preference

This site detects your browser language on first visit. Use the language switcher in the navigation bar to override it; your choice is remembered.