Semantic chunking is a chunking strategy that groups text based on meaning rather than fixed size.
Instead of splitting text after a certain number of tokens, semantic chunking tries to identify:
topic boundaries,
semantic transitions,
and coherent conceptual units.
The goal is simple:
Keep semantically related information together.
