The Markdown Segmenter is a built-in HyperFlow service that intelligently segments markdown-formatted text into smaller, manageable chunks for RAG (Retrieval-Augmented Generation) applications. It preserves document structure, handles images intelligently, and maintains markdown formatting including LaTeX math expressions.
| Property | Value |
|---|---|
| Service Type | hyperflow.segment.markdown |
| Node | Segment Content |
| Content Types | Markdown text (.md files or markdown-formatted content) |
| Service Family | rag.content.segment |
| Parameter | Type | Default | Description |
|---|---|---|---|
minSegSize |
number | 0 | Minimum characters per segment |
maxSegSize |
number | 4000 | Maximum characters per segment |
splitOnHeadings |
boolean | true | Split at markdown headers (#, ##, etc.) |
headingContexts |
boolean | true | Include parent headers for context |
splitOnHorizontalRules |
boolean | false | Split at horizontal rules (---, ***, ___) |
splitTables |
boolean | false | Split at table boundaries |
| Parameter | Type | Default | Description |
|---|---|---|---|
keepImages |
boolean | true | Store images vs remove base64 data |
| Parameter | Type | Default | Description |
|---|---|---|---|
pageNumberDetection |
choice | "None" | Page number detection method |
includePageMarkers |
boolean | false | Add page markers to segments |
keepImages = true