Connect to the Model API
Ask your administrator for your API base URL, model name, and personal API key before connecting a client.
Configure your client
In an OpenAI-compatible client, enter the base URL and model name exactly as provided. For direct API requests, send your key in the Authorization header using the Bearer scheme.
Stream a response
Set "stream": true in a supported chat request to receive the answer as it is generated.
Troubleshoot access
If a request returns 401, check the API key configured in your client. If you receive a model access error, ask your administrator to confirm which models your key can use.