PAPER PENDING

• 45 UX Profiles covering visual, motor, cognitive, and behavioral needs
• Real-time Web Adaptation via Chrome extension
• AI-Powered Personalization using GPT-4 and other LLMs
• Automated Evaluation Suite for accessibility testing
• WCAG 2.2 Aligned profiles and adaptations
• Reproducible Benchmarking without extensive user testing

PersonaLayer consists of two main components:

Real-time web personalization tool that applies AI-generated adaptations based on selected UX profiles.

PL_WebAssist/
├── manifest.json         # Chrome extension configuration
├── sidepanel/           # Main control interface
├── content/             # Page adaptation scripts
├── background/          # Service worker
├── lib/                 # LLM client library
└── assets/              # 45 UX profiles

📖 PL_WebAssist Documentation

Automated testing framework for evaluating web accessibility across profiles and generating comprehensive reports.

PL_WebEval/
├── src/pl_webeval/      # Core evaluation engine
├── scripts/             # Analysis and recovery tools
├── data/                # Test cases and profiles
└── results/             # Test outputs and reports

📖 PL_WebEval Documentation

# 1. Install PL_WebAssist extension
1. Open chrome://extensions/
2. Enable Developer mode
3. Load unpacked → select PersonaLayer_Main/PL_WebAssist
4. Get API key from openrouter.ai
5. Configure in extension side panel

📘 Extension Quick Start Guide

# 1. Install PL_WebEval
cd PersonaLayer_Main/PL_WebEval
pip install -e .

# 2. Set API key
export OPENROUTER_API_KEY="sk-or-..."

# 3. Run evaluation
python run_evaluation.py --testcases data/test_cases.csv

📘 Evaluation Setup Guide

Our profiles are organized into 9 categories:

📖 Complete Profile Documentation

graph TD
    A[User] -->|Selects Profile| B[PL_WebAssist Extension]
    B -->|Captures Page| C[Screenshot + HTML]
    C -->|Sends to| D[LLM API]
    D -->|Generates| E[CSS/JS Adaptations]
    E -->|Applies to| F[Web Page]
    
    G[Researcher] -->|Configures| H[PL_WebEval Suite]
    H -->|Tests| I[Multiple Websites]
    I -->|With| J[45 Profiles × N Models]
    J -->|Generates| K[Reports & Metrics]

• Accessibility Score: 0-100 WCAG compliance rating
• Adaptation Effectiveness: 0-2 scale (none/partial/significant)
• Visual Complexity: Reduced by average 35%
• Cognitive Load: Decreased for 89% of tested profiles

PersonaLayer enables:

• Accessibility Research: Automated testing across diverse user needs
• UX Studies: Profile-based user experience evaluation
• WCAG Compliance: Systematic compliance testing
• AI Personalization: Exploring LLM capabilities for accessibility

• Frontend: Chrome Extension APIs (Manifest V3)
• AI/LLM: OpenRouter API (GPT-4, Claude, Gemini)
• Evaluation: Python, Playwright, pandas
• Analysis: matplotlib, seaborn, LaTeX
• Automation: GitHub Actions, pytest

• Installation Guide
• API Documentation
• Profile Specifications
• Evaluation Methodology
• Contributing Guidelines

We welcome contributions! See our Contributing Guide for:

• Adding new UX profiles
• Improving adaptation algorithms
• Enhancing evaluation metrics
• Bug fixes and optimizations

If you use PersonaLayer in your research, please cite:

@article{XXX,
  title={xxxxx},
  author={[Authors]},
  journal={[Journal/Conference]},
  year={2025}
}

This project is licensed under the MIT License - see the LICENSE file for details.

• WCAG 2.2 Guidelines for accessibility standards
• OpenRouter for LLM API infrastructure
• Chrome Extension team for platform support
• All contributors and testers

• Paper
• Demo Video
• Project Website
• Issue Tracker