codebase-graph

CodeGraph

**Understand your codebase at a glance.** CodeGraph parses your source code, builds a knowledge graph of every function, class, and relationship, then lets you explore, analyze, and query it through AI assistants.

The Problem

Modern codebases are complex. When you need to:

• Understand impact — "What breaks if I change this function?"
• Find security issues — "Where is user input flowing to SQL queries?"
• Reduce complexity — "Which functions are too complex and need refactoring?"
• Navigate relationships — "What calls this? What does this depend on?"

...you're usually stuck grep'ing through files, hoping you didn't miss something.

The Solution

CodeGraph builds a **queryable knowledge graph** of your entire codebase:

Your Codebase  →  Tree-sitter Parser  →  FalkorDB Graph  →  Insights

  42 languages        Functions, Classes, Components, Interfaces, Types
  74 file extensions  CALLS, IMPORTS, EXTENDS, IMPLEMENTS, RENDERS edges
                      Complexity metrics per function
                      Knowledge graph with temporal memory
                      Vector search + cross-encoder reranking
                      Git history integration

Quick Start

Option A: FalkorDBLite (Embedded, No Docker)

The fastest way to get started. FalkorDBLite runs an embedded database with zero external infrastructure:

# Prerequisites: redis-server (macOS arm64 or Linux x64)
brew install redis  # macOS

# Clone and install
git clone https://github.com/your-org/codegraph.git
cd codegraph && pnpm install && pnpm build

# Extract your codebase
pnpm codegraph extract ./src

# Search your codebase
pnpm codegraph search "auth"

A `.codegraph/config.json` file is created automatically:

{
  "driver": "falkordblite",
  "databasePath": ".codegraph/falkordb",
  "graphName": "codegraph"
}

Option B: FalkorDB (Docker)

For development or team/enterprise deployments with a shared graph server:

...