Build your first request.

Reference pages are published as each endpoint is signed off against the backend. The quickstart below reflects the current draft.

Browse models

Documentation

Get started in minutes.

Every endpoint documented with a copyable example and the parameters spelled out inline.

Open documentation

Quickstart

Draft

Create a key in the console, export it, and make your first request. Any OpenAI-compatible client works by changing the base URL.

Shell

export ZAITRO_API_KEY="zai_live_..." curl https://api.zaitroapi.com/v1/chat/completions \  -H "Authorization: Bearer $ZAITRO_API_KEY" \  -H "Content-Type: application/json" \  -d '{ "model": "gpt-5.4", "messages": [{"role":"user","content":"Hello"}] }'

Body parameters

modelstringRequired: The model id to route the request to.
messagesarrayRequired: The conversation so far, oldest first.
temperaturenumber: Sampling temperature. Defaults to 1.
max_tokensinteger: Upper bound on tokens generated in the reply.