Geraph
Structural memory for AI agents. Navigate your codebase with surgical precision.
Geraph maps your entire project—files, functions, classes, and relationships—into a semantic knowledge graph that AI agents (like Claude, Antigravity, and Copilot) can query instead of blindly grepping through your code.
Quick Start
Prerequisite: Node.js >=18.14.0 is required.
Run the following commands to install geraph cli, platform specific rules and build the graph:
npm install -g geraph
geraph install claude
geraph scan
Why Geraph?
- Structural Memory: Long-term "architectural intuition" for AI agents.
- Token Efficiency: Compact 2KB reports instead of MBs of raw code.
- AST Precision: Static analysis (tree-sitter) for absolute accuracy.
- History Aware: Integrates Git history to explain the "Why" behind the code.
- 100% Local & Private: No cloud, no telemetry, no code leaves your machine.
MCP Server (Recommended)
Geraph features a fully local Model Context Protocol (MCP) server that operates completely over stdio. Using the MCP server is highly recommended over running terminal CLI commands for LLMs.
For a project-level local setup:
Add the following configuration to your MCP-compatible client (e.g. Cursor or Antigravity IDE):
{
"mcpServers": {
"geraph": {
"command": "geraph",
"args": ["mcp"]
}
}
}
For a global setup:
If you configure the MCP server globally for your IDE, you must tell the server where your project is located. You can do this by setting the cwd field to your project path. If your IDE/platform doesn't support the cwd field, you can pass the project path as an argument instead:
{
"mcpServers": {
"geraph": {
"command": "geraph",
"args": [
"mcp"
],
"cwd": "<path-to-your-project>"
}
}
}
If cwd is not supported:
{
"mcpServers": {
"geraph": {
"command": "geraph",
"args": [
"mcp",
"<path-to-your-project>"
]
}
}
}
Detailed Documentation
For a full guide on workflows, agent integration, and advanced features, visit the Main Project Documentation.