Google Gemini CLI: Installation Guide and Comparison with Claude Code

Prerequisites: Node.js 20+, terminal familiarity, Google account. Time to complete: ~15 minutes. Sources: Google Blog Announcement, GitHub Repository, Official Documentation

What Google Announced

On June 25, 2025, Google released Gemini CLI — an open-source terminal AI agent under the Apache 2.0 license. It brings Gemini 2.5 Pro directly into the command line with a genuinely usable free tier: 1,000 requests per day with a personal Google account.

This matters because terminal coding agents are becoming the primary interface for AI-assisted development. Developers increasingly prefer staying in their terminal rather than switching between browser tabs, desktop apps, or IDE panels.

What Is Gemini CLI?

Gemini CLI is a terminal-first AI agent that provides lightweight access to Google's Gemini models. It is:

• Open source: Apache 2.0 licensed (GitHub)
• Free tier: 60 requests/minute, 1,000 requests/day with a personal Google account
• Terminal-native: Designed for developers who live in the command line
• Extensible: Supports MCP (Model Context Protocol) for custom integrations
• Multimodal: Accepts PDFs, images, and sketches as input

Built-in Tools

• Google Search grounding
• File operations (read, write, edit)
• Shell command execution (with approval)
• Web fetching
• Codebase querying and editing

Context Window

Gemini 2.5 Pro provides a 1 million token context window, enabling analysis of large codebases in a single session.

Installing Gemini CLI

Prerequisites

• Node.js version 20 or higher
• macOS, Linux, or Windows
• Google account (for free tier)

Option 1: Run Without Installation (npx)

npx https://github.com/google-gemini/gemini-cli

Option 2: Install Globally via npm

npm install -g @google/gemini-cli

Option 3: Install via Homebrew (macOS/Linux)

brew install gemini-cli

Option 4: Install via MacPorts

sudo port install gemini-cli

Option 5: Conda Environment

conda create -y -n gemini_env -c conda-forge nodejs
conda activate gemini_env
npm install -g @google/gemini-cli

Verification

gemini --version

Expected output:

gemini-cli x.y.z

Authentication

Free Tier (Personal Google Account)

gemini auth login

This opens a browser window for Google OAuth authentication. No API key required.

Google AI Studio Key (Higher Limits)

gemini auth login --key YOUR_API_KEY

Vertex AI (Enterprise)

gemini auth login --vertex

First Commands

Explain Code

gemini "Explain this function" < src/utils/helper.js

Fix a Bug

gemini "Fix the null pointer exception in auth.js"

Generate Code

gemini "Create a React component for a user profile card with TypeScript"

Analyze Project Structure

gemini "Analyze this codebase and suggest improvements for error handling"

Multimodal Input (Image)

gemini "Convert this wireframe to React code" < wireframe.png

Working With Existing Projects

Laravel Example

cd ~/projects/my-laravel-app
gemini "Find all N+1 query problems in this codebase"
gemini "Generate a migration for a user preferences table with JSON column"

React Example

cd ~/projects/my-react-app
gemini "Refactor this component to use React hooks instead of class components"
gemini "Add TypeScript types to this file" < src/components/Form.js

API Example

gemini "Generate OpenAPI documentation from these route handlers"
gemini "Create curl commands to test all endpoints in this API"

MCP Support

What Is MCP?

MCP (Model Context Protocol) is an open standard for connecting AI agents to external tools and data sources. Think of it as a plugin system for AI agents.

Why It Matters

MCP allows Gemini CLI to connect to:

• Databases (PostgreSQL, MySQL)
• Version control (GitHub, GitLab)
• Cloud services (AWS, GCP, Azure)
• Internal company APIs
• Custom business tools

Real Example: Database Integration

# Install MCP server for PostgreSQL
npm install -g @modelcontextprotocol/server-postgres

# Configure Gemini CLI to use it
gemini mcp add postgres --command "mcp-server-postgres --connection-string postgresql://localhost/mydb"

# Query database naturally
gemini "Show me all users who signed up in the last 30 days"

Business Use Case

A small business can connect Gemini CLI to their Shopify store via MCP:

gemini "What were yesterday's sales? Compare to last week."
gemini "Find products with low inventory and draft restock emails."

Gemini CLI vs Claude Code vs Codex CLI

Advantages

1. Free tier is genuinely usable — 1,000 requests/day covers most individual developers
2. Open source — inspect code, fork, customize, self-host
3. Apache 2.0 license — safe for commercial use
4. MCP extensibility — connect to databases, APIs, internal tools
5. Google Search grounding — model has access to current information
6. 1M token context — analyze large codebases without chunking
7. Multimodal — process images, PDFs, sketches alongside code

Limitations

1. Requires Google account — no anonymous usage
2. Rate limits on free tier — 60/min may throttle heavy workflows
3. Newer tool — smaller community than Claude Code or Copilot
4. Google ecosystem dependency — authentication tied to Google services
5. Limited IDE integration — primarily terminal-based (VS Code extension in development)
6. Enterprise features cost — Vertex AI required for organization-wide deployment
7. Shell command approval — manual confirmation required for destructive operations

Who Should Use Gemini CLI?

Use Gemini CLI if:

• You prefer terminal workflows
• You want a free tier that covers daily usage
• You need MCP extensibility for custom tools
• You want open-source software under Apache 2.0
• You work with large codebases requiring 1M token context

Consider Claude Code if:

• You need the strongest reasoning for complex architectural decisions
• You have budget for $20/month subscription
• You prioritize reliability over extensibility
• You work primarily in large, complex codebases

Consider Codex CLI if:

• You are already invested in OpenAI ecosystem
• You need Windows desktop integration
• You want usage-based pricing without subscription
• You use GitHub Copilot and want consistency

Final Thoughts

Gemini CLI represents Google's entry into the terminal coding agent space with a genuinely competitive offering. The combination of open-source code, usable free tier, and MCP extensibility makes it attractive for individual developers and technical founders.

The 1,000 request/day free tier is the standout feature — most developers will not hit this limit during normal usage. For comparison, Claude Code requires a $20/month subscription, and Codex CLI charges per token.

The open-source Apache 2.0 license is significant for businesses. You can inspect the code, fork it, modify it for internal use, and deploy it without license concerns.

Start with the free tier using your Google account. Evaluate whether the rate limits constrain your workflow before considering paid options. For most terminal-based development tasks, Gemini CLI provides sufficient capability at zero cost.

Resources

• Google Blog: Introducing Gemini CLI
• GitHub Repository
• Official Documentation
• MCP Documentation

Last updated: June 1, 2026. Pricing and limits subject to change.