Graphify-Labs/graphify — Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.

What happened

The Graphify-Labs team shared a project on GitHub. It reads a whole codebase and builds a knowledge graph. A knowledge graph is a map of things and how they connect. Graphify uses a /graphify command that AI coding tools can run to build this map and then search it.

Key ideas in plain words

Graphify reads code on your own computer using a method called AST parsing. AST parsing breaks code into its parts, such as functions and classes, the way a grammar teacher breaks a sentence into words. This happens in a fixed, repeatable way, so the same input gives the same map. The article also says every link in the graph comes with an explanation. It does not use a vector store, which is a different kind of search database that many AI tools use.

Key facts

• Works with 4 AI coding tools named in the project: Claude Code, Cursor, Codex, and Gemini CLI.
• Reads code plus docs, SQL schemas, configs, and PDFs.
• Parsing runs locally on your machine.
• Licensed under Apache 2.0, a free open-source license.
• Supports many languages; the page lists 30+ translated READMEs.

Why it matters

Large projects are hard to understand, especially for new team members. A graph that shows how parts connect can help people find things faster. Because it runs locally and explains each link, it may be easier to trust than a black-box search.

What to watch

• How well the map works on very large projects.
• Whether more AI tools are added to the list of supported assistants.
• The paid Graphify Cloud service and how it differs from the free local tool.