Markdown Chunking

MarkdownChunking splits markdown on structure like headings, paragraphs, and sections.

Markdown chunking is a method of splitting markdown based on structure like headers, paragraphs and sections. This is useful when you want to process large markdown documents in smaller, manageable pieces.

ParameterTypeDefaultDescription
chunk_sizeint5000Approximate character target. Long atomic content and overlap can make final chunks larger.
overlapint0The number of characters to overlap between chunks.
split_on_headingsUnion[bool, int]FalseHeading-based splitting. False uses size-based chunking. True splits on all headings (H1-H6). An int from 1 to 6 splits on headings at or above that level, e.g. 2 splits on H1 and H2.