Graphify: turn a codebase into a knowledge graph so Claude stops re-reading files
TL;DR
Graphify builds a knowledge graph from a folder of code, docs, papers and images. Instead of re-reading raw files for every question, Claude Code queries the saved graph. Savings grow with project size — on small repos the main benefit is seeing the structure.
How it works
| Step | Details |
|---|---|
| Code analysis | tree-sitter (AST) + call graph, local, no LLM needed |
| Doc analysis | Docs, PDFs and diagram images go through Claude for concept extraction |
| Relationship labels | EXTRACTED / INFERRED / AMBIGUOUS — know what’s certain vs guessed |
| Output | graphify-out/: graph.html, graph.json, GRAPH_REPORT.md, an Obsidian vault |
Install and use
pip install graphifyy && graphify install
# inside Claude Code:
/graphify .How to put it to work
- Onboarding an unfamiliar codebase: open graph.html to see central modules and clusters before reading code.
- Large repos, many questions: that’s where token savings are real; a handful of files barely changes anything.
- Keep it fresh: use the post-commit hook to rebuild.
- Note: the ~70× figure is the project’s own benchmark; docs and images are still sent to Claude when building the graph.
Further reading: Graphify on ClaudeLog