# Chunky

An open-source toolkit for building RAG pipelines that converts PDFs to Markdown, cleans the result, and lets a user visually inspect and compare chunking strategies before indexing.

Chunky is a local, open-source workspace for preparing documents for retrieval-augmented generation, built on the observation that most RAG failures start before retrieval, with broken tables, scrambled layouts, or chunks that look fine in code but fail once placed in context. It combines PDF-to-Markdown conversion through several engines, Markdown cleanup, chunk visualization, and LLM-powered enrichment in one workflow, so a document can be inspected and fixed before it reaches a vector store.

It supports comparing chunking strategies from several popular chunking libraries side by side on the same document with configurable size and overlap, saving and reloading chunk sets tied to the exact Markdown revision that produced them, and generating context-aware titles, summaries, keywords, and retrieval questions for each chunk through an enrichment step. It runs locally or through Docker, and supports batch processing of multiple documents from its sidebar.

Website: https://github.com/GiovanniPasq/chunky