Antigravity
Multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable AI assistant.Workspace template + MCP server for Claude Code, Codex CLI, Cursor & Windsurf.
What is Antigravity?
Antigravity is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable ai assistant.workspace template + mcp server for claude code, codex cli, cursor & windsurf.
Multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable AI assistant.Workspace template + MCP server for Claude Code, Codex CLI, Cursor & Windsurf.
This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Multi-agent knowledge engine (ag-refresh / ag-ask) that turn
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx antigravity-workspace-templateConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Antigravity
Antigravity is a multi-agent knowledge engine that turns any codebase into a queryable AI assistant. It provides two core commands — ag-refresh to build an incremental knowledge base from your project files, and ag-ask to route natural language questions through that knowledge base. It integrates as a workspace template and MCP server with Claude Code, Codex CLI, Cursor, and Windsurf, making it useful for teams that need fast, context-aware answers about large or unfamiliar codebases.
Prerequisites
- Python 3.10+ installed (the CLI and engine are Python packages)
- An MCP-compatible client such as Claude Code or Cursor
- An API key for a supported LLM provider (OpenAI, DeepSeek, Groq, DashScope, Ollama, or any OpenAI-compatible endpoint)
- pip 23.0+ or the uv package runner
- Git, to clone or install from the GitHub repository
Install the CLI and engine packages
Install both the Antigravity CLI and the knowledge engine from the GitHub repository using pip. The CLI provides the ag-* commands and the engine handles embedding and retrieval.
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=cli"
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=engine"Run first-time setup
Run ag-setup inside your project directory. It will prompt you to choose an LLM provider and write a .env file with your API credentials and model settings. Supported providers include OpenAI, DeepSeek, Groq, DashScope, NVIDIA NIM, Ollama, and any OpenAI-compatible endpoint.
ag-setupBuild the project knowledge base
Run ag-refresh to index your codebase. Antigravity scans the project files, generates embeddings, and stores a queryable knowledge base locally. Run this again whenever your codebase changes to update the index incrementally.
ag-refreshQuery the codebase
Use ag-ask to ask natural language questions about your codebase. The engine routes the question through the indexed knowledge base and returns grounded answers with relevant source context.
ag-ask "How does the authentication middleware work?"Start the MCP server for Claude Code
Start the ag-mcp server pointing at your project workspace. This exposes the ask_project and refresh_project tools to any MCP-compatible client such as Claude Code, allowing the AI to query your codebase directly during a session.
ag-mcp --workspace /path/to/your/projectConfigure Claude Code to use the MCP server
Add the Antigravity MCP server to your Claude Code configuration so it is available as a plugin. In Claude Code you can also install via the plugin marketplace.
claude mcp add antigravity -- ag-mcp --workspace /path/to/your/projectAntigravity Examples
Client configuration
Add Antigravity as an MCP server in your claude_desktop_config.json, pointing ag-mcp at the workspace you want to query.
{
"mcpServers": {
"antigravity": {
"command": "ag-mcp",
"args": ["--workspace", "/path/to/your/project"],
"env": {
"OPENAI_API_KEY": "your_openai_api_key",
"OPENAI_MODEL": "gpt-4o"
}
}
}
}Prompts to try
Once connected, use these prompts in your MCP client to explore your codebase through Antigravity.
- "Explain how the database connection pooling is implemented in this project"
- "Where is the rate limiting logic and how does it work?"
- "What does the UserService class do and which modules depend on it?"
- "Refresh the knowledge base and then summarize what changed in the auth module"Troubleshooting Antigravity
ag-ask returns vague or incorrect answers
Run ag-refresh again to ensure the knowledge base is up to date. If the project is very large, check AG_RETRIEVAL_MODE in your .env — set it to 'full' for more thorough context retrieval at the cost of speed.
MCP server not visible in Claude Code after configuration
Verify that ag-mcp is on your PATH by running 'which ag-mcp'. If not found, ensure the CLI package installed correctly and that the Python bin directory is in your PATH. Restart Claude Code after updating the config.
LLM provider authentication errors during ag-setup
Open the .env file created by ag-setup and verify OPENAI_API_KEY and OPENAI_BASE_URL match your provider's credentials. For non-OpenAI providers like Groq or DeepSeek, set OPENAI_BASE_URL to their OpenAI-compatible endpoint URL.
Frequently Asked Questions about Antigravity
What is Antigravity?
Antigravity is a Model Context Protocol (MCP) server that multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable ai assistant.workspace template + mcp server for claude code, codex cli, cursor & windsurf. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Antigravity?
Follow the installation instructions on the Antigravity GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Antigravity?
Antigravity works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Antigravity free to use?
Yes, Antigravity is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Antigravity Alternatives — Similar Coding Agents Servers
Looking for alternatives to Antigravity? Here are other popular coding agents servers you can use with Claude, Cursor, and VS Code.
Dify
★ 142.2kProduction-ready platform for agentic workflow development.
Ruflo
★ 54.0k🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, self-learning swarm intelligence, RAG integrat
Goose
★ 45.7kan open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
Antigravity Awesome Skills
★ 38.3kInstallable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.
AgentScope
★ 25.5kBuild and run agents you can see, understand and trust.
Serena
★ 24.5kA coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.
Browse More Coding Agents MCP Servers
Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Antigravity in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Antigravity?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.