1. Prepare the environment
Keep the key only in an environment variable or a secrets store. Browser code must not contain it or write it to logs.
The integration base URL is https://cicora.ai/api/v1. Each example below appends its route to that base.
First request
bash
export CICORA_API_KEY=replace_with_your_key
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": "user", "content": "Explain this API in one sentence."}]
}'2. Compose the request
- Use the exact model identifier returned by the catalogue.
- Send messages in conversational order: system, user, assistant, and tool when needed.
- Start with lower variance, then tune parameters for the task.
3. Handle the response
Read text from the first available response item and save the request id alongside your own trace id.
Check the status and error structure before retrying. Retry only operations for which your application can prevent duplicate work.