How to set up Claude Code in iTerm2 to run AI projects in one click

The ZDNet columnist described a personal iTerm2 terminal setup on macOS that allows running multiple projects with Claude Code — Anthropic's CLI tool for AI-assisted programming — with a single click, without manually typing commands in each tab.

What's the problem with manual startup

Developers who maintain multiple repositories with Claude Code typically open the terminal, navigate to the required folder using `cd`, and manually run the agent in each project separately. With a dozen active projects, this turns into routine that eats up time before actual work starts.

• Claude Code — Anthropic's CLI agent for writing and refactoring code directly in the terminal
• iTerm2 — popular terminal for macOS with support for profiles and hotkeys
• The setup is built on iTerm2 profiles, which automatically open the required directory and start Claude Code

How the iTerm2 and Claude Code integration works

The solution relies on iTerm2's built-in profiles: each profile is bound to a specific project folder and startup command, so switching between tasks comes down to selecting a profile or pressing an assigned key combination. The terminal itself opens a tab, navigates to the repository directory, and runs Claude Code in it — the agent is immediately ready to accept prompts for that project.

This approach removes repetitive manual actions — typing `cd` and startup commands — and reduces friction when switching between multiple AI coding projects throughout the day.

What this means

As developers use CLI agents like Claude Code in parallel across multiple repositories, customization of the terminal environment becomes part of the daily AI toolkit as much as choosing the model itself.