Choose the AI Model Behind Your Server Edits

PrimeCircuit does not lock you to one brain. You pick the model: a cloud AI provider through a standard API endpoint, a model running on your own PC, or the built-in model, and you can change your choice at any time.

Three ways to connect a model

A cloud AI provider

Enter the endpoint and the API key from a cloud AI provider. This suits large edits and long plugin files, because such models can read more at once. The key stays on your PC, but your prompts go to the provider you chose, so keep private data out of them. The provider may charge separately.

A model on your PC

A model running locally on your PC keeps everything on the machine, so nothing leaves it. This suits private server folders and offline work. Smaller local models can struggle with long files, so give them small, clear jobs such as 1 config change at a time.

The built-in model

The built-in model needs no endpoint and no API key, which makes it a gentle way to try PrimeCircuit on a small config file. For long plugins or tricky logic you may want a bigger model later. Switching takes one visit to the settings.

Check the connection with Test

After you enter an endpoint, press the Test button. It sends a short request, shows the response time and reports the context window, which is how much text the model can read at once. A large plugin file needs a bigger window, so check the number before you start a long job.

Match the model to the job

Use a bigger model for refactors and complex logic, and a smaller one for small edits such as 1 changed value. When a single step is too hard, run_on_model sends just that step to a bigger model while the rest of the chat stays on your usual one.

Quality varies between models. Review the diff whenever you accept a change, and keep Ask first on until you trust how a model handles your files.