@nate_512’s knowledge graph concept has been realized as Graphify, a fully open-source utility that integrates directly with Claude Code and Cursor. The workflow is simple: point the tool at any directory and it automatically constructs a comprehensive knowledge graph without requiring vector databases or configuration files. The resulting output provides a navigable map of every concept, an Obsidian vault complete with backlinks, a wiki generated from an index file, and the ability to query the entire repository in plain English. Its architecture operates in two distinct passes. The initial phase parses code structure locally, bypassing the need for an LLM. In the second phase, subagents work in parallel to extract concepts from documentation and images. Each connection is explicitly labeled as either extracted or inferred, ensuring transparency about what was deduced. This structured approach allows users to pose specific questions such as "What calls this function?", "What connects these two concepts?", and "What are the most important nodes in this project?". By reading the compact graph rather than grepping through raw files, the assistant achieves a 71.5x reduction in tokens per query. This shift represents a fundamentally different paradigm for how AI agents reason across large codebases. Support extends to code written in 13 languages, along with PDFs, Markdown, and images. Installation requires just one command: pip install graphify && graphify install. The project is 100% free, open-source, and part of Y Combinator S26.

@nate_512’s knowledge graph concept has been realized as Graphify, a fully open-source utility that integrates directly with Claude Code and Cursor. The workflow is simple: point the tool at any directory and it automatically constructs a comprehensive knowledge graph without requiring vector databases or configuration files. The resulting output provides a navigable map of every concept, an Obsidian vault complete with backlinks, a wiki generated from an index file, and the ability to query the entire repository in plain English. Its architecture operates in two distinct passes. The initial phase parses code structure locally, bypassing the need for an LLM. In the second phase, subagents work in parallel to extract concepts from documentation and images. Each connection is explicitly labeled as either extracted or inferred, ensuring transparency about what was deduced. This structured approach allows users to pose specific questions such as "What calls this function?", "What connects these two concepts?", and "What are the most important nodes in this project?". By reading the compact graph rather than grepping through raw files, the assistant achieves a 71.5x reduction in tokens per query. This shift represents a fundamentally different paradigm for how AI agents reason across large codebases. Support extends to code written in 13 languages, along with PDFs, Markdown, and images. Installation requires just one command: pip install graphify && graphify install. The project is 100% free, open-source, and part of Y Combinator S26.