Graphify is an advanced AI coding assistant skill designed to transform any folder containing code, SQL schemas, R scripts, shell scripts, documentation, papers, images, or videos into a queryable knowledge graph. Its primary purpose is to provide developers and teams with a comprehensive, interactive map of their project, encompassing application code, database schema, and infrastructure in a single unified graph. This enables users to query relationships, explore architecture, and gain insights without manually searching through files.
Graphify integrates seamlessly with a wide range of AI coding platforms, including Claude Code, Codex, OpenCode, Cursor, Gemini CLI, GitHub Copilot CLI, VS Code Copilot Chat, Aider, Amp, OpenClaw, Factory Droid, Trae, Hermes, Kimi Code, Kiro, Pi, and Google Antigravity. By typing a simple command such as "/graphify ." in your AI assistant, Graphify scans the entire project directory and generates three key outputs: an interactive HTML graph (graph.html), a detailed highlights report (GRAPH_REPORT.md), and a full graph in JSON format (graph.json). These files allow users to visually explore connections, review key concepts and surprising links, and query the graph directly for specific information.
The tool supports extraction from a vast array of file types, including code in over 30 languages, configuration files, markdown and other documentation formats, Office documents, Google Workspace files, PDFs, images, and videos. Code extraction is performed locally using tree-sitter for privacy, while other formats are processed via the user's AI assistant. Optional extras can be installed to extend functionality, such as PDF extraction, Office support, video/audio transcription, Neo4j integration, SVG export, and support for various AI backends (OpenAI, Gemini, Bedrock, Ollama, etc.).
Graphify offers flexible installation options, supporting both user-level and project-scoped installs. It is distributed as the official PyPI package "graphifyy" (with a double-y), and can be installed using modern Python package managers like uv or pipx for ease of use and environment isolation. The skill can be registered with your AI assistant, and platform-specific commands are provided to ensure compatibility across different coding environments.
The generated knowledge graph is designed for collaborative workflows. The output directory (graphify-out/) is intended to be committed to version control, allowing all team members to access the same project map. Graphify also provides hooks for automatic graph rebuilding on git commits and merge drivers to handle concurrent updates. The graph includes special features such as identification of "god nodes" (highly connected concepts), detection of surprising connections across files, extraction of design rationale and comments, suggested questions for exploration, and confidence tags indicating the reliability of inferred relationships.
Users can interact with the graph directly via terminal commands, query for specific paths or explanations, merge graphs, and even expose the graph as a server for repeated tool access. Environment variables and configuration options allow for customization and integration with various AI backends and cloud services. Privacy is a core consideration, with code and media processed locally and sensitive content only sent to AI assistants as needed.
Overall, Graphify serves as a powerful tool for understanding, navigating, and collaborating on complex codebases, making project knowledge accessible and actionable through AI-driven graph mapping.
