Accurate PDF table extraction

Extract complex PDF tables into clean, structured data.

Start extracting tablesView extraction documentation
Core capability

Table extraction built for real-world documents

Extract the columns you specify from dense, multi-page PDFs without forcing every document into a rigid template. Use the web interface or integrate the API into your own document workflow.

  • Column-driven extraction for invoices, statements, and reports
  • Multi-line descriptions and continuation-row handling
  • Multi-page table support with source bounding boxes
  • HTML, JSON, CSV, and XML outputs
  • Automatic MSDI analysis or consumer-provided layout results
  • API integration support, including Tungsten TotalAgility
Tungsten TotalAgility

Get the best table extraction results in TotalAgility

Reuse existing MSDI layout results, map extracted columns directly into document tables, and apply the same integration pattern to invoices, purchase orders, statements, or any other document type.

View TotalAgility integration

Redaction

Upload a PDF, describe what to find and redact, review the detected findings and download a generated artifact.

  • Prompt-based sensitive data detection flow
  • Findings table with coordinates
  • Value-only detection that keeps field labels visible
  • Pixel-burned masks with an optional sanitized search layer

Signature Verification

Upload two signature snippets and receive a visual match confidence, decision, and reasoning.

  • Two image upload workflow
  • Match confidence and decision
  • Reasoning with visual observations
  • API support for base64 image values