Source: http://www.poma-ai.com/docs/blog/rag-chunking/

# RAG Chunking Guide — Where to Start

<ByAuthor />

This page is the entry point. Chunking has one job: turn documents into retrieval units that an embedding model can index and an LLM can use without losing what the text meant in context. Most of the difficulty is that the obvious way to do it, cutting text into pieces, throws that context away. The pages below take you from the strategies people use, through the failures they share, to the approach POMA takes instead.

## Start with these docs

- [Which strategy for which document](/learn/chunking/strategy-landscape) — a decision table by document type
- [Common failure modes](/learn/chunking/common-failure-modes) — what still breaks after you pick a boundary rule
- [POMA chunksets](/learn/chunking/chunksets) — changing the retrieval unit instead of the cut point
- [The full chunking guide](/rag-chunking-strategies-text-splitters) — every strategy in depth, chunk size and overlap, the comparison table

## What this guide is meant to answer

- Which chunking strategies are common in modern RAG systems.
- How chunk size and overlap shape retrieval quality and token cost.
- Why most chunking methods still fail in similar ways.
- How POMA chunksets and cheatsheets change the retrieval unit itself.

<Tldr>
For general-purpose use that permits tradeoffs of accuracy and versatility in exchange for lowered compute costs, recursive delimiter chunking is a popular choice. When the stakes are higher, POMA AI chunksets are designed to preserve hierarchy instead of returning isolated text fragments.
</Tldr>

## Recommended path

If you want the quick structural version, go straight to the [Chunking learning section](/learn/chunking/). If you want the big-picture narrative first, use this page as the entry point and then move through the four topic pages above in order.

## Ready to try hierarchical chunking?

- [Try PrimeCut for free](https://console.poma-ai.com/) — upload a document and inspect the chunks
- [PrimeCut product page](https://www.poma-ai.com/products/primecut-rag-ingestion-chunking) — how it works
- [Pricing](https://www.poma-ai.com/pricing) — PrimeCut Adaptive: 1 credit per page, minute, or 1,000 tokens extracted (whichever is more); 1 credit = €0.01, easy jobs cost less

<AuthorBio />