Google Antigravity
Native MCP / AGYAutonomous Agentic Context for Google Antigravity & AGY CLI
Autonomous AI coding agents in Google Antigravity can veer off course or make incorrect library choices without clear architectural boundaries and verified stack context.
Meto acts as the persistent knowledge brain for Antigravity, letting planner agents and subagent trees read your architectural directives and avoid unwanted refactors.
Key Advantages with Google Antigravity
Step-by-Step Configuration Guide
Configure Antigravity MCP Server
Add Meto to your Antigravity global MCP server configurations in ~/.gemini/antigravity/mcp/ or via the agy CLI settings.
Test with Agent Planner
In Antigravity chat, trigger an audit: 'Read my Meto profile to verify my architectural conventions.'
Use my Meto MCP profile — summarize who I am, what I'm building, and my tech stack.{
"mcpServers": {
"meto": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://www.metoctx.com/api/mcp/your-username",
"--header",
"Authorization: Bearer meto_mcp_your_personal_token"
]
}
}
}Frequently Asked Questions
How does Antigravity access Meto tools and resources?
Antigravity lazily loads Meto's MCP tools (get_profile, get_project_context, list_directives) and resources, calling them only when relevant to minimize token overhead.
Can I use Meto with the Antigravity Python SDK?
Yes. Any Antigravity agent or workflow created via the Python SDK can connect to Meto's standard MCP server protocol.
Ready to give Google Antigravity persistent memory?
Create your Meto profile in 2 minutes. Fill in your stack once, and never repeat your setup again.