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 it solves

Graphify transforms a codebase and its associated documentation (PDFs, images, videos) into a queryable knowledge graph. This allows developers to understand complex projects by querying relationships and tracing paths between concepts rather than manually searching through files using grep or reading raw source code.

How it works

Graphify uses a combination of deterministic parsing and semantic analysis:

  • Code Parsing: Uses tree-sitter AST to parse code across approximately 40 languages locally. This process is deterministic and does not require an LLM, ensuring code remains on the machine.
  • Semantic Pass: For non-code files like PDFs, images, and videos, it uses a configured AI model or API key to perform a semantic pass.
  • Graph Construction: It creates a graph where nodes are concepts and edges are connections (e.g., calls, imports, inherits). Edges are tagged as EXTRACTED (explicitly found in source) or INFERRED (derived by the tool).
  • Integration: It integrates as a "skill" for various AI coding assistants (like Claude Code, Cursor, and GitHub Copilot), allowing the assistant to query the graph instead of reading files one by one.

Who it’s for

Developers and engineers who need to navigate and understand large, complex codebases, as well as those using AI coding assistants who want to provide their assistants with a structured, high-level architectural understanding of their project.

Highlights

  • Local-first code mapping: Code is parsed locally via tree-sitter AST without needing an LLM.
  • Multi-modal support: Maps code, docs, PDFs, images, and video/audio into a single unified graph.
  • Queryable interface: Provides commands to explain concepts, find the shortest path between two nodes, or perform plain-language queries.
  • Visualizations: Generates an interactive graph.html for browsing the codebase as a force-directed graph.
  • Broad assistant compatibility: Supports over 20 AI coding assistants via custom skill installations.
  • Confidence tagging: Distinguishes between explicitly extracted and inferred relationships.