Choose your AI model

Hostcraft Desk runs on your Windows desktop to help you set up static hosting. You pick the brain that powers the agent, and you control where the data goes during every configuration step.

Select your processing engine

Cloud AI provider

Connect a standard API endpoint to use a remote model. Your prompts and file text go to that provider. You retain full control over which endpoint receives your data.

Local PC model

Run a model directly on your Windows machine. The model runs on your PC so prompt text is not sent to a provider. This keeps your configuration secrets and server details on your PC. Note that SSH to your server, web search, http_get and update checks still use the network.

Built-in model

Use the built-in model, which is ready to use without your own endpoint. This option requires no external configuration or API keys and provides a straightforward starting point for basic hosting setups. You can switch to other engines whenever your needs change.

Verify your connection

Press the Test button to check your chosen endpoint. The tool displays the response time and the context window size. This quick check confirms that your model is reachable and ready to process commands for your hosting tasks.

Match the model to the task

Send a difficult step to a bigger model using the run_on_model skill. Use a smaller model for simple edits or quick checks. This strategy balances performance with cost and privacy. You can switch models between different parts of a plan.

Quality varies between different models and providers. Always review the diff in Ask first mode before accepting changes. This habit ensures you understand every modification the agent makes to your configuration files or server settings.