RAG Tools
RAG Chunk Calculator
Calculate chunk counts from document size, chunk size, and overlap.
Estimated chunks
111
Effective chunk size
450
Overlap percentage
10%
Estimated embeddings
111
How it works
RAG systems split documents into overlapping chunks so retrieval can return focused, relevant passages. The overlap helps preserve context across chunk boundaries at the cost of some duplicated content.
Formula
effective step = chunk size − overlap
estimated chunks =
1 if total tokens ≤ chunk size
ceil((total tokens − chunk size) / step) + 1 otherwiseFrequently asked questions
What is the 'right' chunk size?
There isn't one universally correct chunk size — it depends on your documents, embedding model, and retrieval strategy. Smaller chunks improve retrieval precision but increase chunk count and cost; larger chunks reduce count but can dilute relevance.
Why does overlap need to be smaller than chunk size?
Overlap determines how far each chunk advances past the last. If overlap is greater than or equal to chunk size, the effective step becomes zero or negative and chunking never progresses through the document.