Terminal

AI Leash includes an interactive terminal alongside your conversation with an agent. Use it for commands you want to run yourself, or to watch commands that an agent starts while it works on your project.

The terminal is a real shell, so it supports the commands, prompts, and interactive programs you normally use in your project. It is different from the agent’s shell tool, which lets an agent run a command and return its output in the conversation. See Tools for more about that separate feature.

Open a terminal

1. Open the tab picker with the + button in the side panel.
2. Choose Terminal.
3. Type a command and press Enter.

The terminal opens in the project you are currently working in. You can use it to inspect files, install dependencies, run tests, start a development server, check version-control status, or perform any other task supported by your usual command-line tools.

You can interact with a running command just as you would in a normal terminal. For example, press Ctrl+C to stop a process that is running in the foreground. The terminal adjusts to the available space as you resize the side panel. If the shell itself exits, the tab shows a dimmed [process exited] line; close the tab and open a new terminal to start again.

Terminal text uses the code font from Settings > Preferences > Fonts; see Fonts.

Use multiple terminal tabs

Each time you choose Terminal from the tab picker, AI Leash opens another independent terminal tab. This is useful when you need to keep several processes running at once, such as:

• a development server in one tab;
• tests, logs, or a file search in another;
• a separate shell for quick commands.

Switch between terminals by selecting their tabs. Close a terminal with the x on its tab when you are finished with that shell and its running processes.

Run a command from the chat

Shell code blocks in the conversation have an Open in a new terminal button. It opens a new terminal tab with the command already typed at the prompt, without running it, so you can read it, edit it if needed, and press Enter when you are ready.

Watch an Action

Actions are named commands that you can save and run repeatedly, such as npm run dev, a test command, or a project-specific script. They are useful for commands that you and the agent use often, especially commands that stay running in the background.

Open the Actions tab from the tab picker to:

• add an Action by giving it a name and command;
• run or stop a saved Action;
• edit or delete an Action;
• open the Action’s terminal output.

Selecting an Action opens its own terminal tab. When the Action is running, that tab shows its live output, and you can interact with the process there when it accepts input. If the agent starts an Action, you can open the same Action tab to follow along without losing the output that has already been produced. After the Action stops or exits, the tab keeps showing that run’s output, and opening the tab later still shows it. When you run the Action again, the tab follows the new run and clears the previous output as soon as the new run starts.

Everyday use with an agent

The terminal is a convenient place to verify or complement the agent’s work:

• run a test or build after the agent makes changes;
• inspect the current project state while discussing the next step;
• keep a development server available while the agent edits files;
• view the output of a long-running Action without putting it in the chat;
• run a one-off command yourself when you want direct control.

Terminal sessions belong to the project and conversation you are working in, so you can keep your command-line work next to the relevant agent conversation.