pixrep – Visual encoding of codebases for efficient LLM processing

Summary: pixrep converts large code repositories into a structured, hierarchical PDF, enabling multimodal large language models to process codebases efficiently without excessive token use. It generates directory trees, indexes, UML diagrams, call graphs, and lint heatmaps to preserve code structure and context.

What it does

pixrep processes a project folder into a single PDF containing a directory tree, index, auto-generated UML and call graphs for Python, JavaScript, and TypeScript, plus lint heatmaps. It filters out git folders, binaries, and bloat automatically.

Who it's for

Developers working with large codebases who want to provide full repository context to LLMs without token limits or loss of structure.

Why it matters

It prevents token waste and loss of code architecture by visually encoding repositories, allowing LLMs to analyze complete codebases accurately and efficiently.