Connect a Model for Speed Analysis

Decade Web connects to a cloud API, a local PC model, or the built-in option. You pick the provider, test the response time, and in Ask first you review each diff before accepting changes to your site files.

Choose Your Model Source

Cloud AI Provider

Connect Decade Web to a standard API endpoint using your own credentials. The prompt and file text travel to that provider for processing. You retain full control over which cloud service handles your data. This option works well when you need high capacity for complex refactors or large context windows during heavy speed analysis tasks.

Local PC Model

Run a model directly on your Windows machine using Decade Web. With a local model, prompts and file text stay on your PC; web fetches and optional network features still connect out. This setup keeps chats, working folders and API keys on your PC.

Built-in Model

Use the built-in model that is ready to use without your own endpoint. Approval mode controls actions.

Verify Your Connection

Press the Test button to check your chosen endpoint. The interface displays the response time and context window details. This quick check confirms that Decade Web can communicate with your selected model before you begin any speed analysis or file edits on your site.

Select the Right Model Size

Send a hard step to a bigger model using run_on_model for complex logic. Use a smaller model for simple edits or quick grep searches. You can switch between models based on the complexity of the speed issue you are currently investigating.

Quality varies between models, so review the diff carefully. Use Ask first to pause before actions that change something. The agent shows the exact file change. You decide whether to accept the modification or request a different approach for your speed optimization.