Antigravity CLI and Claude Code do the same basic job. You open a terminal in your project folder, describe what you want in plain English, and the agent reads files, proposes edits and runs commands. The big differences are who makes them, what you pay to start, and which AI models you get.
The short version
| Antigravity CLI | Claude Code | |
|---|---|---|
| Made by | Anthropic | |
| What it is | Terminal agent (agy) | Terminal agent (plus IDE and desktop integrations) |
| Open source | Not stated in Google's docs | No |
| Free plan | Yes, an Individuals plan with basic weekly rate limits | No. Needs a paid Claude plan or API credits |
| Sign-in | Personal Google account, Google Cloud, or Gemini API key | Claude account, Console account, or a cloud provider |
| Models | Google's model list includes Gemini and some Claude models | Claude models |
| MCP servers | Yes | Yes |
| Project instructions | AGENTS.md or GEMINI.md | CLAUDE.md, plus skills |
| Skills, hooks, subagents | Yes | Yes |
| Install | One-line installer for macOS, Linux and Windows | Native installer for macOS, Linux and Windows |
Prices and plans change. Check Google's pricing page and Anthropic's pricing page rather than trusting a table (including this one).
A note on cost
Google's pricing page lists an Individuals plan at $0 a month with basic weekly rate limits, and larger limits on Google AI Pro and Ultra. Google doesn't publish numbers for those limits in its docs, so we can't say how far the free plan goes. Rate limits are the thing to watch if you plan to use it every day.
Claude Code needs a paid Claude plan (Pro, Max, Ultra, Team or Enterprise), a Claude Console account with pay-as-you-go credits, or access through Amazon Bedrock, Google Cloud or Microsoft Foundry. The free claude.ai plan doesn't include it.
How they feel to use
Day to day they're very similar. You type a request, the agent works out which files matter, shows what it wants to change and asks before anything risky. Both can also run a single prompt from a script instead of chatting. In Antigravity CLI that's agy -p "your prompt".
Antigravity CLI shows an inline diff for each edit by default and has working modes you can cycle through, including a planning mode. It also shares its engine with Google's Antigravity desktop app.
Where Antigravity CLI is stronger
- A free way in. You can sign in with a personal Google account and start on the free plan.
- Google's ecosystem. It works with Google Cloud sign-in, and with Gemini API keys.
- Reads
AGENTS.md. One instructions file can be shared with other tools.
Where Claude Code is stronger
- A clearer paid path. The plans and credits are well documented by Anthropic.
- Claude models by default. If you prefer how Claude writes and explains code, this is where you get it.
- Established extras. Skills, hooks, subagents and
CLAUDE.mdmemory are central, well-documented features. Antigravity CLI has versions of these too, but it is newer, and Google says it doesn't have full parity with the Gemini CLI it replaces.
What they share
Both support MCP (Model Context Protocol, a standard way to plug extra tools such as GitHub or a database into an AI). Both read a plain-text file in your project for standing instructions. Both ask permission before changing things by default.
Who should pick what
- You want to try an agent without paying first: Antigravity CLI.
- You already pay for Claude: Claude Code.
- You want your instructions in
CLAUDE.mdwith skills and hooks: Claude Code. - You've never used a terminal: either will feel unfamiliar. Start with our Claude Code install guide or the Antigravity CLI getting started guide.
- You can't decide: try both on one small, real task. They don't conflict.
Whichever you choose, set up a safety net first. Our git safety net guide shows how to undo anything the AI gets wrong.
Our verdict
There's no universal winner. Antigravity CLI is the easy, free-to-start option from Google. Claude Code is the more established, feature-rich one with a clear paid path. See also Antigravity CLI vs Gemini CLI and Claude Code vs OpenAI Codex.