Document Splitting

Split large documents into smaller, focused sections.

What is Splitting?

Document splitting divides large documents into logical sections:

• By page ranges (manual)
• By AI-detected boundaries (automatic)
• By structure (headings, chapters)

Use cases:

• Split large reports into chapters
• Separate multi-page forms
• Extract specific sections for processing
• Organize long documents

Splitting Documents

From the Dashboard

1. Go to Split in the sidebar
2. Select a document
3. Choose splitting method:
   • Manual: Define page ranges
   • AI: Let DocLD detect sections
4. Review suggested splits
5. Apply

AI-Powered Splitting

DocLD can automatically detect logical section boundaries:

1. Upload or select a document
2. Click Analyze
3. Review detected sections
4. Adjust if needed
5. Click Split

AI detection looks for:

• Chapter headings
• Section titles
• Page breaks
• Table of contents entries
• Structural patterns

Splitting Methods

Manual Split

Define exact page ranges:

{
  "method": "manual",
  "sections": [
    {
      "name": "Introduction",
      "page_start": 1,
      "page_end": 5
    },
    {
      "name": "Chapter 1",
      "page_start": 6,
      "page_end": 20
    },
    {
      "name": "Appendix",
      "page_start": 21,
      "page_end": 30
    }
  ]
}

AI Split

Let AI detect boundaries:

{
  "method": "ai",
  "instructions": "Split by chapter headings and major sections"
}

AI returns suggested sections with confidence scores.

Page-Based Split

Split by fixed page counts:

{
  "method": "page",
  "pages_per_section": 10
}

Split Configurations

Save reusable split configurations:

Creating a Config

{
  "name": "Legal Document Splitter",
  "description": "Split legal documents by article",
  "instructions": "Detect Article and Section headings",
  "settings": {
    "method": "ai",
    "min_section_pages": 2,
    "preserve_headers": true
  }
}

Using a Config

Apply a saved configuration:

1. Select a document
2. Click Apply Config
3. Choose your configuration
4. Review and apply

Saved configs appear in the Configurations tab on the Split page. You can also save the current configuration from that tab for reuse. Via the API, saved configs and inline config can include advanced options (e.g. settings.splitMode, settings.minSectionPages); see the Split API for the full shape.

Section Types

After Splitting

Each section becomes a new document that you can:

• Extract - Run extraction on specific sections
• Chat - Ask questions about specific sections
• Organize - Add sections to different knowledge bases
• Export - Download individual sections

Section Relationships

Sections maintain a relationship to the parent:

{
  "id": "section-uuid",
  "parent_document_id": "original-uuid",
  "section_name": "Chapter 1",
  "page_start": 6,
  "page_end": 20
}

Workflow Integration

Add splitting to workflows:

{
  "type": "split",
  "config": {
    "method": "ai",
    "instructions": "Split by section headings"
  }
}

This enables automatic document splitting as part of processing pipelines.

Best Practices

1. Analyze first - Use AI analysis before splitting
2. Review sections - Verify detected boundaries
3. Name clearly - Use descriptive section names
4. Save configs - Reuse configurations for similar documents
5. Process sections - Extract from specific sections for accuracy

API Reference

See the Split API for programmatic access.