cicora.ai
English
API specificationText generation

Chat Completions

Create multi-turn conversations in the familiar messages format and select a model for each request.

Integration reference: choose parameters and available capabilities from your account API settings and the model catalogue for your environment.

Messages and roles

The messages array carries context as an ordered sequence. System establishes task rules, user adds a request, and assistant and tool preserve execution history.

Do not build history by concatenating strings: retain the role and original order of every message.

Chat completion
bash
curl --request POST https://cicora.ai/api/v1/chat/completions \
  --header "Authorization: Bearer $CICORA_API_KEY" \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "openai/gpt-5.6-sol",
    "messages": [
      {"role": "system", "content": "Be concise and cite assumptions."},
      {"role": "user", "content": "Draft a release checklist."}
    ],
    "temperature": 0.2
  }'

Control the output

  • model identifies a model family and variant.
  • temperature, top_p, and an output limit control response variance and size.
  • Send only parameters listed as supported for the selected model.

Reliable loop

Separate a transport failure from a model response with a finish_reason. A refusal or context boundary is interface data, not a reason to substitute a result.

When your data-retention rules allow it, store the source request body and normalized response without credentials.

Continue with