cicora.ai
English
API specificationAnswers

Developer FAQ

Short practical answers about keys, base URLs, models, cost, and integration diagnosis.

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

Where should I start?

Create a separate key, configure the OpenAI-style base URL https://cicora.ai/api/v1, and run GET /models. Then send a minimal chat request with a selected model identifier.

For a Messages SDK, use the gateway root https://cicora.ai/api. On a 404, compare the final URL with the client protocol: one appends a route after /v1, the other appends /v1/messages.

A safe first diagnostic request
bash
curl --include --request GET https://cicora.ai/api/v1/models \
  --header "Authorization: Bearer $CICORA_API_KEY" \
  --header 'Accept: application/json'

Frequent solutions

  • Do not store a key in client code or publish it in an issue or screenshot.
  • Do not treat a price example as current pricing: read catalogue and response metadata.
  • Do not transfer parameters or capabilities between models without checking the model.

When support is needed

Collect timestamp, method, URL without query secrets, HTTP status, request id, and a de-identified error body.

Do not attach a key, full prompts containing restricted data, or files unless they are necessary to reproduce the issue.

Continue with