Table of Contents

Property DocumentVisionParser

Namespace
LMKit.Retrieval
Assembly
LM-Kit.NET.dll

DocumentVisionParser

Gets or sets an optional vision-based analyzer for document understanding.

public VlmOcr DocumentVisionParser { get; set; }

Property Value

VlmOcr

A VlmOcr instance for visual document analysis, or null (default).

Examples

using var chat = new PdfChat(chatModel, embeddingModel);

// Configure vision-based document understanding
LM visionModel = LM.LoadFromModelID("gemma4:e4b");
chat.DocumentVisionParser = new VlmOcr(visionModel);

Remarks

When set, enables visual analysis of document pages, interpreting layout, structure, and content meaning beyond raw text extraction.

Used when PageProcessingMode is set to DocumentUnderstanding, or when set to Auto and visual analysis would benefit a given page.

If not set, document processing falls back to text extraction methods only, optionally using OcrEngine for image-based content.

See Also
Share