Gemini CLI

Overview

Gemini CLI is an open-source AI agent that lets you use Gemini directly from the terminal. It is built for developers who work mainly on the command line and focuses on shortening the path from prompt to model.

It includes tools such as grounding with Google Search, file operations, shell command execution and web page fetching. It supports MCP (Model Context Protocol) so you can connect custom capabilities, and its multimodal features let it build apps from PDFs, images and sketches.

You can start with a personal Google account within a free usage allowance, which makes it easy for terminal-focused developers to try. Releases are split into preview, stable and nightly channels.

What it is used for

• Query and edit large codebases.
• Build new apps from PDFs, images and sketches.
• Debug problems and find their causes using natural language.
• Automate operational tasks such as looking up pull requests or handling complex rebases.
• Extend it by connecting new capabilities through MCP servers.

Pricing and license

Apache 2.0 license. Free usage allowance with a personal Google account

Getting started

1. To try it without installing, run npx @google/gemini-cli .
2. For a global install, use npm install -g @google/gemini-cli or brew install gemini-cli .
3. Run Gemini CLI in the terminal and log in with your personal Google account.
4. If needed, connect MCP servers to extend its capabilities.

Links

Official site ↗Repository ↗

Related terms and companies

Coding agent AI agent MCP Tool use Multimodal Context window Google

Sources

• google-gemini/gemini-cli README

Compiled from public sources (official pages, READMEs). Pricing and features change, so check the official site.