Antigravity CLI — Getting Started

Antigravity CLI (agy) replaces Gemini CLI for individual consumers. This guide covers installation, authentication, and your first session.

Installation

Quick install (recommended)

curl -fsSL https://antigravity.google/cli/install.sh | bash

The script detects your OS and architecture, downloads the Go binary, verifies the checksum, and installs to ~/.local/bin/agy.

npm

npm install -g @google/antigravity-cli

Homebrew (macOS/Linux)

brew install antigravity-cli

Verify

agy --version

If the command isn't found, add ~/.local/bin to your PATH:

echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc

Authentication

Antigravity CLI supports three authentication methods:

Option 1: OAuth (recommended)

agy

Choose "Sign in with Google" and follow the browser flow. Your token is cached in the system keyring.

Option 2: API Key

export ANTIGRAVITY_API_KEY="your-api-key"
agy

Get a free API key at aistudio.google.com/apikey. The free tier supports 1,000 requests/day.

Option 3: Enterprise (Vertex AI / Code Assist)

export GOOGLE_CLOUD_PROJECT="your-project-id"
agy

First Session

cd your-project
agy

You'll see the Antigravity CLI prompt:

⚡ Antigravity CLI
  Model: Gemini 3.5 Flash (High)
  Context: 1,048,576 tokens
  Sandbox: workspace

  Type /help for commands.

>

Essential Commands

| Command | What it does | |---|---|---| | /help | Show all commands | | /model | Switch AI model | | /config | Open configuration settings | | /clear | Clear conversation | | /mcp | Manage MCP servers | | /skills | Manage installed skills and plugins | | /agents | List and monitor active agents |

Quick Examples

# Analyze your codebase
> Explain the architecture of this project

# Generate code
> Create a REST API endpoint for user authentication

# Debug
> Why is this test failing?

# Refactor
> Rename `User` to `Account` across all files

Model Selection

Switch models during a session:

/model

Available models:

Or set via flag:

agy --model "Claude Sonnet 4.6 (Thinking)"

Next Steps

• Prompt Engineering — Write effective prompts
• Configuration Reference — Settings, MCP, plugins
• Migration Guide — If migrating from Gemini CLI
• Patterns — Subagents, hooks, async tasks