lead · github
Graphify turns your codebase into a queryable knowledge graph for Claude and Cursor
Parse code, docs, SQL schemas, and PDFs into a deterministic graph. Feed it to your AI assistant. No vector store hallucinations. No weeks of manual context engineering.
Graphify-Labs/graphify ↗Here's the problem: when you hand Claude or Cursor a codebase, it's just text. The AI has no real understanding of how things connect. You end up manually explaining the architecture, schema relationships, and config dependencies over and over.
Graphify solves this by reading your entire codebase, documentation, database schemas, and config files, then mapping out the actual connections between them as a graph. Think of it like creating a living blueprint. You can then inject that graph directly into Claude Code, Cursor, or Gemini CLI. The AI understands structure, not just tokens. It queries the graph deterministically, so you get consistent answers grounded in your actual code, not hallucinated guesses.
Share kit
Graphify: turn legacy codebases into queryable knowledge graphs for Claude
Stop copy-pasting context. Graphify parses your code, docs, SQL schemas, and PDFs into a deterministic graph that Claude and Cursor can query directly. Local parsing, no vector store, no hallucinations. Free to use via CLI or Cursor skill. Saves weeks of manual prompt engineering.
Your codebase is already a graph. Graphify just makes it queryable. Parse code + docs + SQL schemas + PDFs into a deterministic knowledge graph. Feed it to Claude Code or Cursor. No vector stores. No hallucinations. No weeks of context engineering. https://github.com/Graphify-Labs/graphify
Legacy codebase giving your AI assistant a hard time? Graphify solves a real bottleneck: making existing systems actually queryable. Local deterministic parsing. Code, docs, SQL schemas, PDFs all mapped into a knowledge graph. Inject into Claude, Cursor, or Gemini CLI. No vector store. No hallucinations. Free. Open source. Worth trying today if you're shipping with AI context at scale. https://github.com/Graphify-Labs/graphify
Just shipped a workflow that's saved me weeks of context engineering. Built a parser that turns your entire codebase, docs, SQL schemas, and PDFs into a queryable knowledge graph. No vector store. No hallucinations. Just deterministic AST parsing and every edge explained. Loaded it into Claude Code this morning. Asked it questions about a legacy system I haven't touched in two years. Got back answers grounded in actual code structure, not statistical guesses. It's a /graphify skill for Claude, Cursor, Codex, and Gemini CLI. One command. Your codebase becomes a graph. Then you query it like you'd query a database. Built for anyone who's ever had to manually copy-paste context into an AI to make it useful.
graphify just crossed a milestone. local deterministic AST parsing that turns your entire codebase + docs + schemas into a queryable knowledge graph. no vector store. no hallucinations. works with claude, cursor, codex, gemini cli. one command. your legacy system becomes queryable. https://github.com/Graphify-Labs/graphify