Auto-Tag PDF (Docling IBM)

Start automating accessibility with IBM’s Docling AI — a 100% free, 100% local solution that transforms PDF remediation with enterprise-grade accuracy, zero cloud dependency, and complete data privacy.

PDFix now integrates with Docling – IBM Research’s open-source AI toolkit. This powerful combination delivers professional-grade auto-tagging completely free and fully local – no cloud uploads, no subscriptions, no data privacy concerns.

Meet PDFix Desktop + Docling: AI Auto-Tagging That Works

• 100% Free: Open-source Docling model + PDFix Desktop integration
• Runs Locally: Process sensitive documents on your own machine via Docker
• IBM Research AI: Trained on 81,000 manually labeled pages from diverse document types
• Lightning Fast: Auto-tag complete documents in seconds, not hours
• Batch Processing: Process entire folders with identical workflows
• Enterprise Accuracy: Layout detection, table recognition, heading identification

How It Works: Choose Your Path

For PDFix Desktop Users

Perfect for accessibility professionals, document specialists, and teams who prefer visual workflows.

• Download PDFix Desktop
• 🐳 Install Docker Desktop First!
• Open PDFix Desktop and pull the docling docker via Action Manager (one-time set up)
• Upload your PDF → External Actions → AutoTag (Docling)

For PDFix SDK Users

Perfect for software developers, enterprises with custom workflows, and organizations needing programmatic integration.

• Automated Pipeline Integration
  • Build the Docling action into your document processing pipelines, web applications, or enterprise systems using PDFix SDK.
• Resources for SDK Integration:
  • 📦 Docker Hub: hub.docker.com/r/pdfix/pdf-accessibility-docling
  • 💻 GitHub Repository: github.com/pdfix/action-pdf-accessibility-docling-docker
• SDK Benefits
  • Programmatic control over entire workflow
  • Integrate with existing systems
  • Batch process unlimited volumes
  • Template generation and reuse
  • Multi-threaded processing

Local Processing

• Complete data control
• No internet required after setup
• Unlimited processing
• No per-document costs
• Works offline

The Technology Behind It

• Docling Layout Model
  • Uses RT-DETR object detection to analyze document layouts, identifying text blocks, images, tables, and captions with near-human accuracy.
• PDFix SDK Integration
  • The Dockerized solution connects Docling AI with PDFix proven remediation engine, enabling seamless auto-tagging workflows right inside PDFix Desktop
• Template System
  • You can either use the action directly for automatic tagging or add an intermediate step by generating a layout template, which you can review and adjust if needed. Read more about Layout Templates.

Resources

• Getting Started Guide: https://pdfix.net/user-guide-external-actions/
• GitHub Repository: https://github.com/pdfix/action-pdf-accessibility-docling-docker
• Docker: https://hub.docker.com/r/pdfix/pdf-accessibility-docling

Actions