Getting Started with 100X Code: From Install to Your First Agent

1. Install the CLI

100X Code ships as a single command-line tool. On macOS or Linux, install it with:

The installer uses bun when it is available and falls back to npm, so you can also install it directly with bun install -g 100xprompt-cli or npm install -g 100xprompt-cli. When it finishes, the 100xprompt command is on your PATH.

2. Point It at a Model

100X Code runs against an engine you control. If you are on managed Flash, connect with the endpoint from your dashboard. If you run Pro on your own hardware, point the CLI at your inference server. Either way, no prompt leaves your perimeter. New to the trade-off? Read Pro vs Flash.

3. Run Your First Task

Change into a project and start the agent:

Then describe an outcome in plain language, like "explain this codebase" or "add input validation to the signup handler and write a test for it." 100X Code will plan the steps, make the edits, run whatever it needs to, and show you the result. This plan, act, observe loop is the heart of how the CLI works.

4. Give It Your Tools and Memory

The agent gets dramatically more useful once it can reach your systems. Connect your tracker, database, or wiki over MCP, and let Chitta remember your team's conventions across sessions. From here on, the assistant is not just capable, it knows your project.

Where to Go Next

The full command reference and configuration live in the documentation. If you are coming from a cloud tool, the migration guide maps the path. And if you want the big picture first, start with the platform overview.