1. Create an account, then a key.
Create a Codimixa account or sign in to an existing one. Review the account’s applicable model rates, balance and billing requirements before sending requests. Create an API key in the console, then store it as CODIMIXA_API_KEY in your local environment. Keep it out of source control and browser-side code.
2. Send a small Chat Completions request.
This Bash example uses a model that supports Chat Completions. Replace YOUR_MODEL_ID with its exact public ID before running the request.
curl https://api.codimixa.com/v1/chat/completions \
-H "Authorization: Bearer $CODIMIXA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_MODEL_ID",
"messages": [{"role": "user", "content": "Reply with one short greeting."}]
}'Use Responses for a GPT agent workflow.
GPT-6.1 Sol requires Responses for tool calling. Start with this text request, then add the tools your application runs. The API base URL remains https://api.codimixa.com/v1.
curl https://api.codimixa.com/v1/responses \
-H "Authorization: Bearer $CODIMIXA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-6.1-sol",
"input": "Reply with one short greeting.",
"max_output_tokens": 256
}'3. Inspect the result and usage.
Check the HTTP result, the response content and any returned usage fields. Compare the request with its account usage record. If the request fails, read the error before trying again.
4. Connect your application or agent.
Use the SDK guide for Python or JavaScript, or follow the setup guide for your client. Check streaming and tool behavior with the selected model when your application needs them.
A FEW DETAILS
Before you continue
Which price applies to my request?
Use the price shown for your selected model and account. Selected model rates start at 0.3× official public API pricing.