Connect an AI model for your Lua code

Codelab Station works with a cloud AI provider through a standard API endpoint, with a model running on your PC, or with the built-in model. Pick one. Then press Test, and you can start on your addon or server files.

Three ways to connect a model

A cloud AI provider

A cloud AI provider suits hard logic and long files, because larger models can keep more of your code in view. Enter its API endpoint and key in Codelab Station. Your prompts, including the code you attach, go to that provider, so use it only for files you may share. Keys stay on your PC.

A model on your PC

A model running on your own PC suits private code and work without a connection, and you point Codelab Station at its local endpoint. Nothing leaves the machine. That makes it the choice for server settings you want to keep to yourself. Speed and quality depend on your hardware and on the model you load.

The built-in model

The built-in model is the starting point when you have no API key and no local model yet. Test it on a copy. Prompts go only to the model endpoint you pick, so choose that endpoint on purpose before you attach private code.

What the Test button shows

Press Test first. The button checks that Codelab Station can reach your endpoint and shows the response time and the context window, which tells you how much code fits into one request. Run it again after every change of provider, key or model name.

Picking a model for the job

Use a bigger model for refactors and complex logic, such as a gamemode that spreads its state over several files, and a smaller one for small edits like a changed chat message. run_on_model sends one hard step to a bigger model while the rest of the task stays on your usual choice. Size follows the task.

Quality varies between models. Read the diff, use Ask first for anything you cannot judge at a glance, and test on gm_construct before you trust a change.