Connect your AI model

CodeSpar runs on your Windows desktop and needs an AI brain to understand your code. You choose how it thinks by picking a provider, a local model, or the built-in option.

Choose your model

Cloud provider

Use a standard API endpoint to send your prompts to a cloud AI provider. Your file text leaves your PC to reach that provider. The Test button shows the response time and context window.

Local model

Run a model directly on your PC. With a local model nothing leaves your machine.

Built-in model

Use the built-in model that comes ready to use without your own endpoint.

Check your connection

Press the Test button to verify your model endpoint. The tool shows the response time and the context window. This helps you confirm that the model is reachable and ready to process your prompts for the SQL injection and XSS testing tasks.

Pick the right model

Send a hard step to a bigger model using run_on_model for complex logic. Use a smaller model for small edits to save resources. The agent orchestrates these tasks. You can switch models for different parts of your testing workflow.

Quality varies between models, so review the diff carefully. Use Ask first mode to pause before actions that change something. This lets you accept or reject the change. You stay in control of every edit made to your staging copy.