Accessible PDF AI

Production-grade accessible PDF system with Cloudflare Workers, .NET 10, and HPC/SLURM integration.

Status

🚧 Current Phase: Project Scaffold
🎯 Next: MVP Development (see roadmap)

Why Use This?

This project aims to automate PDF accessibility remediation using modern AI models.

• Batch Processing: Enable processing of large document collections without manual review
• Cost Efficient: Leverage existing HPC infrastructure for cost-effective processing
• Deployment Flexibility: Cloud, hybrid, or fully on-premise with HPC integration
• Modern AI: Uses LayoutLMv3, BLIP-2, and other state-of-the-art models for document understanding and alt-text generation
• WCAG 2.1 AA Compliance: Automated enforcement, not just checking. Easily upgrade compliance standards.

Designed for UC Davis document remediation needs.

Learn more about the rationale →

Repository Structure

• client/ – React 19 + Vite frontend with TanStack Router and Query
• hpc_runner/ – SLURM job runner (heavy ML + PDF, Python)
• server/ – .NET 10 Web API bridge to HPC cluster
• server.core/ – Domain models and shared logic
• tests/ – .NET test project with xUnit
• workers/ – Cloudflare Workers (R2 + Queues, TypeScript)

Quick Start

# Copy environment template
cp .env.example .env

# Setup all development environments
just setup

# Run CI checks (linting + tests)
just ci

# Start all services with Docker
just dev

# Before committing (includes coverage)
just commit-check

# See all available commands
just help

Dev Container

For a consistent development environment, use the provided Dev Container configuration:

VS Code:

1. Install the Dev Containers extension
2. Open this project in VS Code
3. Click "Reopen in Container" when prompted

GitHub Codespaces:

1. Click the green "Code" button on GitHub
2. Select the "Codespaces" tab
3. Click "Create codespace on [branch]"

The dev container includes Node.js, .NET SDK, Python/uv, and all required VS Code extensions. See .devcontainer/README.md for details.

Documentation

• Architecture - System design and data flow
• Development Guide - Setup and workflows
• MVP Roadmap - Path to functional prototype
• System Design - Technical decisions
• Testing Strategy - Test coverage and approach
• Why This Project? - Problem statement and rationale

Operations

• Monitoring Setup - Prometheus and Grafana configuration
• Metrics Deployment - Cloudflare Worker metrics infrastructure
• Security Setup - Authentication and secure configuration

MVP Progress

Track development: MVP v0.1 Milestone

10 issues across 5 phases:

1. Infrastructure (R2, Queue, Database)
2. Backend (Controller, SLURM)
3. Processing (Mock PDF analysis)
4. Frontend (Upload, Status)
5. Validation (E2E tests)