How to Install Gemini CLI on Linux

A comprehensive step-by-step setup and deployment guide

Table of Contents

• Step 1: Install Prerequisites (Node.js)
• Step 2: Install Gemini CLI
• Step 3: Launch and Authenticate
• Quick Usage Examples

Installing the official Google Gemini CLI (an open-source AI agent for your terminal) on Linux is quick and straightforward. Because it is a Node.js-based application, you just need a compatible Node environment configured on your system.

Step 1: Install Prerequisites (Node.js)

The Gemini CLI requires Node.js version 20.0.0 or higher.

If you are using a Debian/Ubuntu-based system, the default package manager versions might be outdated. Using NVM (Node Version Manager) or the NodeSource repository is highly recommended.

Option A: Using NVM (Recommended)

1. Run the NVM installer:

   curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.3/install.sh | bash
   

2. Restart your terminal or source your profile configuration:

   source ~/.bashrc
   

   (Or use source ~/.zshrc depending on your active shell)

3. Install and switch to Node 20+:

   nvm install 20
   nvm use 20
   

Option B: Using Ubuntu/Debian Package Manager

Alternatively, you can pull the LTS binaries directly from NodeSource:

sudo apt update && sudo apt install -y curl
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt-get install -y nodejs

Step 2: Install Gemini CLI

Once Node.js and npm are properly configured, you can install the CLI package globally on your system.

sudo npm install -g @google/gemini-cli

Alternative (No-Install Method): If you just want to evaluate or try out the CLI tool without permanently installing it to your system paths, you can leverage npx to execute it instantly on demand:

npx @google/gemini-cli

Step 3: Launch and Authenticate

To initialize the setup and start the CLI runtime environment, simply execute:

gemini

Authentication Options

Upon your very first launch sequence, the CLI tool will prompt you to choose an interface theme and complete your authentication. You have two main pathways:

• Option 1: Sign in with Google (OAuth) [Highly Recommended] — Selecting this opens a standard browser session to authenticate securely via your personal Google account. This structural tier grants a generous free allocation (typically allowing up to 1,000 requests per day) without requiring active token management.
• Option 2: Gemini API Key — If you prefer a headless interface or standard backend token architecture, fetch an API key from Google AI Studio and export it directly to your target shell pipeline environment:

  export GEMINI_API_KEY="your_actual_api_key_here"
  

Quick Usage Examples

Once completely configured, you can query and leverage the AI agent directly inside your pipeline infrastructure:

• Interactive Mode: Launch the persistent chat session:

  gemini
  

• Single Prompt (Non-Interactive): Run queries directly and capture stdout:

  gemini "Explain how to optimize this bash script"
  

• Specify a Specific Model: Override default operational models:

  gemini -m gemini-2.5-flash
  

• Include Local File Context & Directories: Feed structural directory code context explicitly to the agent:

  gemini --include-directories ./src,./docs
  

Google Gemini CLI Documentation | 2026