Graphify: Turn Your Codebase Into a Queryable Knowledge Graph

Graphify is a skill for AI coding assistants that does one precise, useful thing: you point it at a folder of code, SQL schemas, scripts, docs, images, or videos, and it transforms everything into a queryable knowledge graph. Visual, queryable, zero frustration when you actually need to find something.

How it works? Dead simple: you install the Python package (graphifyy—double y because developers), hook it up with uv or pipx, then type /graphify in your IDE. Done. You get three files: an interactive HTML graph (clickable, filterable), a Markdown report with key concepts and unexpected connections, and a graph.json you can query without re-reading anything.

The genius bit: you don't want a summary of your project. You want to navigate it. You want Claude or Cursor to tell you "module X talks to database Y through Z," not "go find it yourself." The graph does exactly that—it gives your AI a real, structured understanding of your codebase that it can query in real time.

Compatibility? It covers the bases: Claude Code (Linux/Mac/Windows), Codex, OpenCode, Cursor, Gemini CLI, GitHub Copilot CLI, VS Code Copilot Chat, Aider, and more. It's not locked to one vendor—it's a modular skill that works wherever your AI lives.

Minimal prerequisites: Python 3.10+, uv or pipx to install. On Mac, brew handles it. On Windows, winget. It's not rocket science.

One detail worth noting: the repo's got 60k GitHub stars, 116 open issues, 179 pull requests, and supports 23 languages. That means it actually works, gets maintained, and real people use it.