AI-powered PDF data extraction tool with visual verification and confidence

Precision PDF is an open-source document processing platform that extracts structured data from PDFs while showing you exactly where every piece of data comes from. Built with Next.js 15, Convex, and Clerk authentication.

• 🔍 Visual Data Verification - See exactly where extracted data comes from in the original PDF
• ⚡ Real-time Processing - Live updates as documents are processed
• 📊 Smart Table Recognition - Automatic table detection and CSV export
• 📄 Multiple Export Formats - JSON, CSV, DOCX, Markdown, Text, XLSX
• 🏥 Document Type Support - Invoices, medical records, bank statements, forms
• 📱 Multi-page Documents - Handle complex documents with multiple pages
• 🔌 API Access - Full REST API for developers
• 🎯 Interactive Demo - Try 8 real examples without signing up

This repository is currently configured for easy local development with ALL AUTHENTICATION AND SECURITY FEATURES DISABLED.

For production deployment, you MUST:

• Re-enable authentication in middleware.ts
• Configure all environment variables properly
• Follow the Security Configuration Guide

See Security Documentation for complete details.

• Node.js (Latest LTS recommended)
• pnpm package manager
• Convex CLI (npm install -g convex)

# Clone the repository
git clone https://github.com/yourusername/precision-pdf.git
cd precision-pdf

# Install dependencies
pnpm install

# Set up environment variables
cp .env.example .env.local

# Initialize Convex (creates a new deployment)
npx convex dev

# Start the development server
pnpm run dev

Your app will be running at http://localhost:3000

Note: The FastAPI processing service is optional for local development. Example documents work without it.

graph TD
    A[Next.js Frontend] --> B[Convex Backend]
    A --> C[API Routes]
    C --> D[FastAPI Service]
    D --> E[Landing AI]
    B --> F[Document Storage]
    B --> G[User Management]
    A --> H[Clerk Auth - DISABLED]
    C --> I[Stripe Payments]

Core Components:

• Frontend: Next.js 15 with App Router and Tailwind CSS
• Backend: Convex for real-time database and serverless functions
• Authentication: Clerk (currently disabled for local development)
• Processing: External FastAPI service with Landing AI
• Payments: Stripe integration
• UI Components: shadcn/ui component library

# Start development servers (both frontend and backend)
pnpm run dev

# Run only frontend (Next.js)
pnpm run dev:frontend

# Run only backend (Convex)
pnpm run dev:backend

# Build for production
pnpm run build

# Run tests
pnpm run test              # Unit tests with Vitest
pnpm run pw:test          # E2E tests with Playwright
pnpm run pw:test:ui       # Playwright UI mode

# Linting and formatting
pnpm run lint

Copy .env.example to .env.local and configure:

# Core Services (Required)
NEXT_PUBLIC_CONVEX_URL="https://your-deployment.convex.cloud"
NEXT_PUBLIC_APP_URL="http://localhost:3000"

# Authentication (Clerk) - Currently disabled
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY="pk_test_your-clerk-key"
CLERK_SECRET_KEY="sk_test_your-clerk-secret"

# Document Processing (Optional for local dev)
FAST_API_URL="http://localhost:8000"
FAST_API_SECRET_KEY="your-secret-key"

# Payments (Stripe) - Optional for local dev
STRIPE_PUBLISHABLE_KEY="pk_test_your-stripe-key"
STRIPE_SECRET_KEY="sk_test_your-stripe-secret"

See Environment Variables Guide for complete reference.

1. Convex - Backend database and serverless functions

   • Sign up at convex.dev
   • Free tier available

2. FastAPI Service (Optional for local development)

   • Repository: precision_pdf_fast_api
   • Handles PDF processing with Landing AI
   • Can run locally or deploy to Render

3. Clerk - Authentication (currently disabled)
4. Stripe - Payment processing
5. Landing AI - Document processing AI
6. Sentry - Error monitoring

The project includes comprehensive testing infrastructure:

# Unit Tests (Vitest)
pnpm run test          # Run once
pnpm run test:watch    # Watch mode
pnpm run test:ui       # UI interface

# E2E Tests (Playwright)
pnpm run pw:test       # Headless
pnpm run pw:test:ui    # UI mode
pnpm run pw:test:debug # Debug mode

Currently no tests are implemented, but infrastructure is ready. See Testing Guide.

• Next.js 15 - React framework with App Router
• React 19 - UI library
• Tailwind CSS - Utility-first styling
• shadcn/ui - Component library
• TypeScript - Type safety

• Convex - Real-time database and serverless functions
• Clerk - Authentication (currently disabled)
• Stripe - Payment processing

• FastAPI - Document processing service
• Landing AI - AI-powered document extraction

• Vercel - Frontend hosting
• Render - FastAPI hosting
• Sentry - Error monitoring
• PostHog - Analytics

The app includes 8 pre-processed example documents:

• 📧 Invoice
• 🏦 Bank Statements (2)
• 🏥 Medical Reports (2)
• 📑 Medical Journal Article
• 🏠 Mortgage Application
• 📋 Settlement Statement

Examples are stored in /public/examples/ and can be explored without authentication.

We welcome contributions! Please see our Contributing Guide for details on:

• Code style and standards
• Development workflow
• Pull request process
• Issue reporting

# Fork the repo and clone your fork
git clone https://github.com/yourusername/precision-pdf.git

# Create a feature branch
git checkout -b feature/your-feature-name

# Make your changes and test
pnpm run test
pnpm run lint

# Submit a pull request

"User not authenticated" errors in development:

• This is expected since authentication is disabled
• Check the security configuration guide

Documents not processing:

• Ensure FastAPI service is running
• Check environment variable configuration
• See Troubleshooting Guide

Build errors:

• Ensure you're using the latest Node.js LTS
• Delete node_modules and run pnpm install

• ✅ Core document processing
• ✅ Visual verification interface
• ✅ Multiple export formats
• ✅ Real-time processing updates
• ⚠️ Authentication (disabled for local dev)
• ⚠️ Testing (infrastructure ready)
• 🔄 Documentation (in progress)

This project is open source. License details coming soon.

• Documentation: Browse /docs folder
• Issues: GitHub Issues
• Discussions: GitHub Discussions

For questions about this project:

• GitHub: @robertguss
• Website: precisionpdf.com

⭐ Star this repository if you find it useful! ⭐