Chunking
Slicing long documents into bite-size passages so retrieval can fetch just the relevant piece, not the whole file.
Think of it like
Cutting a book into index cards so you can pull the one paragraph you need.
Example
A 40-page manual gets split into ~300-word chunks, each embedded and searchable on its own.
How it actually works
Chunk too big and retrieval drags in irrelevant text that dilutes the answer; too small and you sever the context a passage needs to make sense. A little overlap between chunks keeps ideas from being cut mid-thought. It’s unglamorous, but chunking quality quietly decides RAG quality.
For product teams
A boring knob with outsized impact on answer quality — worth tuning early.
For engineers
Segmenting docs into overlapping passages sized for the embedding model and context budget.
Related
- Retrieval — Prepares documents for retrieval.
- Embeddings — Each chunk becomes an embedding.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome