Built-in Segmenter

Overview

The Built-in segmenter is HyperFlow's comprehensive content segmentation service, offering specialized methods for Text, PDF, HTML, and Markdown content. It provides powerful, format-aware segmentation strategies optimized for different document types and use cases.

Service Information

Property	Value
Service Name	Built-in segmenter
Status	Enabled
Compatible Nodes	Segment Content

Supported Content Types

Text: Simple text segmentation with sliding windows, per-file splitting, or pattern-based separation
PDF: Fast, structured text extraction with intelligent paragraph detection and recursive splitting
HTML: Web content segmentation with three different extraction methods and optional CSS selector-based page destructuring.
Markdown: Structure-aware, multi-modal segmentation that preserves document hierarchy, tables, math formula, etc.

Configuration Parameters

Required Parameters

Parameter	Type	Default	Description
Segmentation source	choice	-	Content type selection (PDF, HTML, Text, or Markdown)
Service	choice	varies	The segmentation method for the selected content type

Content Type: Text

Note: The text segmentation options in this service provide fast, basic text splitting. For more sophisticated text segmentation with semantic awareness or recursive splitting, consider using the Langchain segmenter.

Sliding Windows

Creates overlapping text segments of fixed size, useful for maintaining context across boundaries.

Parameter	Type	Default	Description
Window size	number	1024	Size of each text window in characters
Overlap	number	64	Number of overlapping characters between windows

How it works: Moves a fixed-size window through the text with specified overlap, creating segments that share context at boundaries.

Best for: General text where consistent chunk sizes are important, embeddings generation, or when context preservation is critical.