Graphify: a map of your repo you can query

Graphify turns your codebase into a graph you can query instead of grep. The code part runs with no LLM at all.

Type /graphify in your coding assistant and it maps the project into a knowledge graph. Per the README, code is parsed with tree-sitter, so that part is deterministic and nothing leaves your machine. Docs, PDFs, images and video go through your assistant's model for a semantic pass.

Two details I like. Every edge in the graph is tagged EXTRACTED (it's explicit in the source) or INFERRED (Graphify worked it out), so you can see what was read and what was guessed. And it isn't a vector index. There are no embeddings and no vector store, just a graph you traverse: ask a question, trace the path between two things, or explain one concept.

It has 128k stars and an Apache 2.0 license. There's also a hosted product, Graphify Cloud, with a 14-day free trial, and the README's '80% less token spend' headline is about that one. I'd measure the open-source version on your own repo before believing any number.

Install is short: uv tool install graphifyy, then graphify install, then /graphify in the repo.

When your agent explores a codebase, do you trust it more with grep or with a map?

Sources

#ClaudeCode#KnowledgeGraph#OpenSource#AI

New notes go out on LinkedIn first. Building something with AI? We can help.

Diagram card: Graphify: a map of your repo you can query