graphify hits 450K downloads in 26 days

/graphify is a skill for AI coding assistants that reads codebases, documents, images, and audio·video and turns them into a knowledge graph. The author revealed that the project recorded 450,000+ downloads in 26 days and about 40,000 GitHub stars.

• Code and SQL structures are extracted via AST, while documents, images, and transcripts are analyzed in parallel using Claude subagents.
• The input range is broad, covering code, PDF, Markdown, screenshots, diagrams, whiteboard photos, and audio·video.
• Results are saved as graph.html, graph.json, and a cache directory, allowing reuse across different sessions.
• Relationships are categorized as EXTRACTED, INFERRED, or AMBIGUOUS, separating directly confirmed content from inferences.
• The graph is clustered using Leiden community detection, without using separate embeddings or a vector DB.
• It can be installed on multiple agents/IDEs including Claude Code, Codex, Cursor, Gemini CLI, and GitHub Copilot CLI.