Content Comparison

The Unique platform offers the possibility to use different supports multiple services for ingesting PDF documents:

• Default Unique Ingestion Service

• Docling

• Microsoft Document Intelligence (MDI in the following)

• MDI with Image Content Extraction

Each service can parse structured PDFs with a single-column layout and extract simple tables. However, their capabilities vary when handling more complex documents:

• Image-based PDFs: Scanned or printed PDFs lack structured content, requiring OCR techniques for extraction.

• Multi-Column Layout: PDFs with multiple columns, charts, tables, and text need pre-trained layout detection models to identify page elements and preserve logical content flow.

• Complex Tables Detection: Extracting tables with merged cells, missing borders, or checkmarks requires specialized AI models to recognize different table components.

• Image Content Extraction: Many PDFs contain unstructured visual elements like charts, logos, or photos. AI models with image-to-text capabilities are needed to extract this content in a searchable form.

• On-Prem Deployment: The service can operate in a closed environment without internet access.

General recommendation:

• On-prem customers should deploy Prem Customers: Use Docling for PDF ingestion, as the Unique’s Default ingestion service is not capable of processing Default Unique Ingestion Service lacks efficient support for multi-column layouts in a performant way.

• Cloud customers should use Customers: Use MDI as a the default ingestion service, as it extracts information from PDFs with a higher precision than Docling such as tables that are missing grid lines. Customers that have PDF documents with provides higher accuracy than Docling, particularly for tables without grid lines.

  • If PDFs contain charts or table-like structures,

  the

  • MDI with Image Content Extraction

  should be used as it will make almost all the content that can be contained in a PDF document

  • is recommended for making all document content searchable and accessible to

  a

  • language

  model.

  • models.

- fully supported 🟡 - partially supported - not supported

Ingestion service	Capabilities					Performance	Additional costs
	Structured PDFs	Unstructured PDFs	One-Column LayoutImage-based PDFs	Multi-Column LayoutLayouts	Extracts Complex Tables Detection			Detects Images	Extracts Image Content Extraction	On-prem deploymentPrem Deployment
Default									10-15s per page	None
Docling	🟡				🟡			10-20s per page	Azure infra Costs
MDsI		MDI							10-20s per page	1.6 cents per page
MDI with Image Content Extraction									20-30s per page	3 cents per page Assumption: 1.6 cents for MDI 1.4 cents for 5k tokens (vision model GPT4o) per image per page (assuming 1 image per page)

...