PDF Accessibility

AI-powered PDF accessibility automation with WCAG compliance levels and intelligent workflow orchestration.
Join our community

Overview

This node automates PDF accessibility enhancement and compliance checking with WCAG 2.1 standards using AI-powered workflows. It supports multiple processing modes, including fully automated intelligent processing, analysis-only, remediation-only, and customizable workflows where specific accessibility operations can be selected.

Common scenarios include:

  • Automatically analyzing PDFs to identify accessibility issues and applying fixes to meet legal or organizational standards.
  • Generating detailed accessibility reports for auditing purposes.
  • Improving document structure, alt-text for images, heading optimization, table accessibility, link descriptions, metadata, reading order, and color contrast.
  • Integrating into document processing pipelines to ensure all PDFs are accessible before distribution.

Practical examples:

  • A publishing company uses the node to automatically remediate PDFs before release, ensuring compliance with accessibility laws.
  • An educational institution generates accessibility reports on course materials to identify improvements.
  • A developer customizes the workflow to only enhance image alt-text and metadata for marketing PDFs.

Properties

Name Meaning
Processing Mode How the node processes the PDF:
- Intelligent Auto-Processing (AI determines optimal workflow)
- Analysis Only (accessibility analysis without changes)
- Remediation Only (apply fixes to analyzed PDF)
- Custom Workflow (select specific operations)
WCAG Compliance Level Target WCAG 2.1 compliance level:
- Level A (Minimum)
- Level AA (Standard, recommended)
- Level AAA (Enhanced, comprehensive)
LLM Provider Large Language Model provider for AI tasks:
- Auto-Select Best Provider
- Anthropic (Claude)
- OpenAI (GPT)
- Google (Gemini)
PDF Input Method How to provide the PDF file:
- Binary Data (from previous node)
- File Path (local system)
- Download from URL (including Google Drive links)
- Base64 Data
Binary Property Name Name of binary property containing PDF data (shown if input method is Binary Data)
File Path Full local file path to the PDF (shown if input method is File Path)
URL URL to download the PDF (shown if input method is Download from URL)
Base64 Data Base64 encoded PDF data (shown if input method is Base64 Data)
Advanced Settings Collection of advanced options:
- Processing Depth: Quick Fixes, Comprehensive, Deep Analysis
- Document Language (e.g., en-US, es, fr)
- Output Format(s): Enhanced PDF, Analysis Report, Compliance Summary, Remediation Log
- Enable Debug Mode
- Show Cost Estimation
Custom Operations Specific accessibility operations to perform when in Custom Workflow mode:
- Document Structure Analysis
- Image Alt-Text Generation
- Heading Optimization
- Table Accessibility
- Link Text Improvement
- Metadata Enhancement
- Reading Order Optimization
- Color Contrast Validation

Output

The node outputs JSON objects containing detailed results depending on the processing mode:

  • Intelligent Auto-Processing:

    • mode: "intelligent_auto"
    • wcagLevel: Target compliance level
    • documentInfo: Metadata about the PDF (file name, page count, size, text length, presence of images/tables/links, language)
    • workflowPlan: Recommended tools and reasoning behind their selection
    • execution: Summary of tools executed, total issues found and fixed, processing time, success status
    • accessibilityReport: Summary, compliance score, recommendations, detailed findings
    • auditReport: Detailed audit report in multiple export formats (JSON, HTML, Markdown, CSV)
    • settings: Advanced settings used
    • timestamp and processingTime
    • version: Node version

  • Analysis Only, Remediation Only, Custom Workflow:

    • Simple JSON messages indicating completion and readiness for future implementation phases.

If errors occur and "continue on fail" is enabled, output includes an error message JSON object.

The node does not output binary data directly but processes PDF content internally.

Dependencies

  • Requires access to AI Large Language Model providers via API keys or credentials configured in n8n for Anthropic, OpenAI, or Google services.
  • Uses internal utility modules for PDF input handling, validation, and accessibility tools management.
  • Supports downloading PDFs from URLs, reading from local files, or receiving binary/base64 data from other nodes.
  • Optional debug mode enables detailed logging for troubleshooting.
  • No direct external environment variables beyond standard API credential configurations.

Troubleshooting

  • Common Issues:

    • Invalid or inaccessible PDF input (wrong file path, broken URL, missing binary data).
    • Missing or misconfigured API credentials for selected LLM provider.
    • Unsupported PDF format or corrupted files causing validation failure.
    • Selecting unknown processing mode triggers an error.

  • Error Messages:

    • "Unknown processing mode: ...": Check that the "Processing Mode" property is set correctly.
    • Errors related to PDF validation indicate problems with the input PDF file. Verify the source and format.
    • API authentication errors suggest missing or invalid credentials; verify API keys in n8n credentials.
    • If the node fails and "Continue On Fail" is off, the workflow will stop; enable it to capture errors as output instead.

  • Resolution Tips:

    • Confirm PDF input method matches provided data.
    • Ensure API credentials for the chosen LLM provider are properly set up.
    • Use debug mode to get detailed logs for diagnosing issues.
    • Validate URLs and file paths externally before use.

Links and References

Discussion