Graphify — Code Knowledge Graph
An AI coding assistant skill for Claude Code that turns any folder of code, SQL schemas, docs, papers, or images into a queryable, interactive knowledge graph.
View on GitHubMultimodal Input
Processes code, PDFs, markdown, screenshots, diagrams, and images in any language.
Interactive HTML Graph
Outputs a visual graph with search, filtering, and clickable nodes you can explore.
Persistent Storage
graph.json lets you query weeks later without reprocessing your entire codebase.
71.5x Token Reduction
Dramatically cuts AI token usage versus reading raw files across large codebases.
Auto-Sync Watch Mode
Watch mode updates graphs in real time as files change; git hooks rebuild after commits.
Multiple Export Formats
Export to HTML, Obsidian vaults, Wikipedia-style wikis, SVG, GraphML, and Neo4j.
Install via pip
Use pipx on macOS if you encounter environment errors.
Run the installer
Windows users: add Python Scripts to PATH if the command isn't found.
Use in Claude Code
Type this inside Claude Code to analyze your current directory.
Optional extras
How It Works
Graphify analyzes your project files using Claude AI, extracts concepts and relationships, and builds a structured knowledge graph stored in graph.json. You can then query it in plain English, export it to multiple formats, or view it as an interactive HTML visualization.
Input
Code, SQL, docs, images, PDFs
Process
Claude AI extracts nodes & edges
Output
graph.json + HTML visualization
Export Formats
Edge relationships are tagged as EXTRACTED, INFERRED, or AMBIGUOUS for full transparency.