Agentic AI Learning Portal

Appearance

🅰️ Antigravity CLI Reference Guide ​

Antigravity CLI is Google's terminal-first AI development platform - the successor to Gemini CLI (retired June 2026). It runs autonomous coding agents, executes shell commands, and orchestrates multi-agent workflows directly from your terminal.

📊 Tool Datasheet ​

MetricDetails
Tool NameAntigravity CLI
CategoryTerminal AI Development Platform
PurposeAutonomous coding agents with deep Google Cloud integration - plan, code, test, debug, and deploy from the terminal.
DeveloperGoogle
LicenseCommercial / Free tier available
PredecessorGemini CLI (retired June 18, 2026)
Websiteantigravity.google

🛠️ Capabilities ​

  1. Autonomous Coding Agents: Agents that plan, write code, run tests, debug failures, and deploy - all from the terminal.
  2. Multi-Agent Orchestration: Decompose tasks across specialized agents working in parallel with fan-out and merge patterns.
  3. Google Cloud Integration: Native deploy to Cloud Run, GKE, Firebase. Deep integration with Vertex AI, Google Workspace, and Gemini models.
  4. Terminal-First UX: Lightweight CLI with streaming agent output. Full shell passthrough for direct command execution.
  5. Model Access: Backed by the Antigravity agent model - 76.2% on Terminal-Bench 2.1, outperforming Gemini 3.1 Pro on coding and agentic benchmarks.

🧭 Step-by-Step Setup ​

1. Installation ​

bash
npm install -g @google/antigravity-cli

2. Authentication ​

bash
antigravity login

Follow the browser-based OAuth flow to authenticate with your Google account.

3. Initialize a Project ​

bash
cd ~/your-project
antigravity init

This creates an .antigravity/ directory with project context and configuration.

4. Run an Agent Task ​

bash
antigravity run "Fix the authentication bug in src/auth.ts and verify tests pass"

The agent searches the codebase, applies edits, runs tests, and reports results - all streaming to your terminal.

5. Multi-Agent Workflow ​

bash
antigravity run --parallel "Build the API, write tests, and update documentation"

Fan-out: the task splits across specialized agents for implementation, testing, and documentation - results merge when all agents complete.

💡 Sample Workflows ​

  1. Code Migration: antigravity run "Migrate all Express routes to FastAPI, update type annotations, and verify endpoints with curl"
  2. Dependency Audit: antigravity run "Audit package.json for vulnerable dependencies, suggest upgrades, and apply safe updates"
  3. Deploy to Cloud Run: antigravity run "Build the Docker image, push to Artifact Registry, and deploy to Cloud Run with the staging tag"