01 — Sign in
Use an available sign-in method. Registration is available only when enabled. Check your balance, group access, and limits before sending a paid request.
02 — Choose a model
Open model details in the catalog. Check the supported endpoint, billing unit, and group conditions. Copy the exact model name.
- Distinguish prices per 1K and per 1M tokens.
- Check input and output rates separately.
03 — Create an API key
Open API keys in the console. Create a named key for your application with appropriate limits and expiry.
- Keep it in a server environment variable or secret store.
- Never publish it in browser code or a repository.
- Revoke and replace an exposed key.
04 — Make the connection
Set the base URL to https://aiapi.ge/v1 when your client requests a base URL. Raw HTTP examples use the full endpoint path. Start with a short, nonsensitive prompt.
05 — Verify the result
Check usage records for the model and charge. Test both success and failure before increasing traffic. A successful HTTP response is not proof that the generated answer is correct.