chunk-engine
Chunking Modes

sentence

A fixed number of sentences per chunk for tight token budgets.

View raw

sentence produces chunks of a fixed number of sentences. When you have a tight token budget and want predictable chunk sizes, this mode gives you consistent, sentence-aligned boundaries.

from py_chunks import get_chunks

chunks = get_chunks("article.docx", mode="sentence", sentences_per_chunk=3)

Parameters

ParameterDefaultNotes
sentences_per_chunk3Sentences per chunk. Must be greater than 0.

content_type

Chunks carry the content_type sentence. Metadata typically records the actual sentence count per chunk (the last chunk may hold fewer).

When to use it

  • Strict token budgets where uniform chunk sizes matter.
  • Models or indexes that assume short, even inputs.
  • Sentence-level highlighting or citation.

On this page