> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lynxhub.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Card launch configuration

> Configure start arguments, pre-launch scripts, window layouts, and address detection for individual cards.

Each card module on the dashboard has its own launch configuration. Use this configuration to manage launch arguments, pre-launch scripts, browser viewports, and custom commands.

To open the launch configuration panel:

1. Locate the card on your dashboard.
2. Click the **3 dots menu** in the top-right corner of the card.
3. Click **Launch Config** from the dropdown menu.

***

## Visual argument manager

The visual argument manager allows you to toggle command-line arguments without editing scripts:

* **Presets**: Choose from pre-configured argument presets or save your own custom presets using the **Save Preset** option.
* **Custom arguments**: Toggle individual argument flags (such as `--use-cpu`, `--lowvram`, or `--xformers`) to control how the local AI model runs on your hardware.
* **Support custom arguments toggle**: You can enable or disable custom visual argument rendering using the **Custom Arguments UI** setting.

***

## Pre-launch configurations

Configure scripts and resources that execute before the main backend process boots.

### Terminal commands

Add commands to prepare your execution environment:

* Use this to activate virtual environments: `conda activate myenv` or `.\venv\Scripts\activate`.
* Use this to set environment variables on Windows: `set CUDA_VISIBLE_DEVICES=0`.
* Use this to set environment variables on Linux/macOS: `export CUDA_VISIBLE_DEVICES=0`.

To add a command:

1. Click **Add Command** under the **Pre-Launch Commands** section.
2. Type the shell command.
3. The host executes these commands in sequence in the terminal tab before launching the module.

### Open files and folders

Automatically open files or directories in native system applications when the card launches:

* Select **Add File** or **Add Folder**.
* Specify the absolute path (such as `D:\AI\outputs` or `D:\AI\config.json`).
* When the card launches, LynxHub opens these folders or configuration files in your default explorer or editor.

***

## Custom run configurations

Override the default startup scripts and configure process settings:

* **Custom launch command**: Enable this to bypass default run commands. Enter your own command line execution string (e.g. `python launch.py --listen --theme dark`).
* **Launch behavior**: Choose the viewport display layout when running the card:
  * **Split View (Default)**: Show terminal stream output on the left and the web browser on the right.
  * **Terminal Only**: Hides the browser and displays only the terminal tab.
  * **Browser Only**: Hides the terminal tab and opens the browser full-screen.
  * **None**: Launches the process invisibly in the background.
* **Browser load delay**: Set the delay (in milliseconds) before the integrated browser starts loading the address. The default is `0ms`.
* **Auto-close terminal on exit**: Enable this to close the terminal tab automatically when the backend process terminates.

***

## Address detection

LynxHub reads the stdout stream from the terminal process to detect when the local web server is ready.

Customize how LynxHub captures this address:

* **Automatic detection**: Uses standard port patterns (such as `http://127.0.0.1:7860`) to redirect the browser.
* **Custom port override**: Force the browser to open a specific address or port.
* **Regex catch pattern**: Set a custom regular expression pattern to parse the console output and extract the local URL.