A practical walkthrough of chunking for retrieval-augmented generation: why chunk boundaries decide retrieval quality, how to implement fixed, recursive, and semantic splitting, and how to verify the results.