ATAPI Twin
Pages

Build

OpenAI-compatible models

Chat completions and the model list, without calling a provider.

The OpenAI template creates POST /v1/chat/completions and GET /v1/models. The reply is deterministic from the prompt. stream: true returns server-sent events and ends with data: [DONE].

A request that includes tools, and a prompt that mentions tool, function, or weather, returns tool_calls. The Faults tab can force tool calls whenever tools are present. The model name, preamble, and chunk timing are on that same tab.

curl -sS https://apitwin.dev/sim/YOUR-SLUG/v1/chat/completions \
  -H 'content-type: application/json' \
  -d '{"messages":[{"role":"user","content":"ping"}]}'