Skip to main contentOnce your model is deployed and running, you can manage it through a dedicated dashboard.
Model Management Interface
The dashboard for your active model is organized into several tabs:
-
Overview: This is your primary control panel, displaying the model’s status (e.g., “Running,” “Deploying”), its unique endpoint URL, and a sample cURL request to help you quickly integrate it into your applications. Your API key is also conveniently displayed here for easy access.
-
Playground: An interactive interface that allows you to send test prompts directly to your model and view the generated output in real-time. This is an excellent tool for quick validation and experimentation without writing any code.
-
Events: A high-level log of significant activities related to your deployment, such as its creation, status changes, and any system-level notifications.
-
Logs: Provides access to the raw, detailed logs from the model’s deployment. This is an invaluable resource for debugging and troubleshooting any operational issues.
Terminating a Model Deployment
If a model is no longer needed, you can terminate its deployment by clicking the Terminate button on its dashboard. This action is permanent and will remove the model and all associated data from the platform.
Crucially, terminating a model deployment does not automatically release the underlying GPU cluster. The GPU instance will remain active and continue to incur charges until you manually release it from the GPUs panel.
Please refer to Step 3 of Getting Started 
