Configuration

Open Limitcode: Open Settings to edit the user settings file. A minimal setup looks like this:

{
    "default_provider": "openai",
    "default_model": "gpt-5.5",
    "api_keys": {
        "openai": ""
    },
    "temperature": "auto",
    "max_tokens": 8192,
    "max_iterations": 50,
    "chat_font_size": "auto",
    "reasoning_effort": "off",
    "show_thoughts": false,
    "auto_save_edits": false
}

Common options

Option	Values	Description
default_provider	Provider name	Provider used for new sessions
default_model	Model name	Model used for new sessions
temperature	"auto" or number	Controls response randomness
max_tokens	"auto" or positive integer	Maximum response length
max_iterations	Positive integer	Maximum tool loop iterations
chat_font_size	"auto" or positive number	Font size for the Limitcode chat view
reasoning_effort	off, low, medium, high	Reasoning level where supported
show_thoughts	true or false	Show model reasoning where supported
auto_save_edits	true or false	Save files automatically after edits

Provider keys can also be configured through Limitcode: Setup Provider API Key. Never commit keys to the repository or paste them into issue reports.

temperature accepts "auto" or a number. max_tokens accepts "auto" or a positive integer. Unsupported models ignore reasoning_effort.

Chat appearance

The chat font follows Sublime Text’s global font size by default. Set a fixed size with "chat_font_size": 11. The chat status bar can also be customized with chat_style, or with a user CSS override at Packages/User/Limitcode/chat.css.