Manifold Companion

A document-processing platform built by the CUNY AI Lab that transforms scanned pages, historical manuscripts, and digital PDFs into publication-ready texts for CUNY's Manifold instance.

Workflows

Image to Markdown — Upload images (JPEG, PNG, TIFF, BMP, WebP, HEIC) or rasterized PDFs. AI-powered OCR extracts text into editable Markdown. Review and correct page by page, then export.

PDF to HTML — Upload a digital PDF and the platform converts it to structured HTML preserving headings, tables, lists, and mathematical formulas (TeX/KaTeX). Edit in a rich-text editor, then export.

Both workflows include AI-powered summaries and translations (40+ languages), Dublin Core metadata, version history, and ZIP export ready for Manifold ingestion.

Tech Stack

• Backend: Express 4, better-sqlite3, express-session
• Frontend: React 18, React Router 6, Tailwind CSS 3, KaTeX
• AI/OCR: AWS Bedrock (Image-to-Markdown pipeline), Google Gemini via OpenRouter (PDF-to-HTML pipeline)
• Infrastructure: SQLite with WAL mode, AWS SES for email notifications

Both AI providers operate under zero data retention policies.

Getting Started

# Install dependencies
npm install

# Configure environment
cp .env.example .env
# Edit .env with your AWS credentials, OpenRouter API key, session secret, etc.

# Development (starts both server and client)
npm run dev

# Production build
npm run build
npm start

Environment Variables

Variable	Description
SESSION_SECRET	Express session secret (>=32 chars in production)
AWS_REGION	AWS region for Bedrock and SES
BEDROCK_OCR_MODEL	Vision model ID for OCR
BEDROCK_TEXT_MODEL	Text model ID for summaries/translations
OPENROUTER_API_KEY	API key for OpenRouter
ADMIN_EMAIL / ADMIN_PASSWORD	Seeds initial admin user
SES_FROM_EMAIL	SES verified sender address
APP_URL	Public app URL for email links

Features

• OCR: AI vision model extracts text from scans, photos, and handwritten pages
• Rich editing: Side-by-side page image and text editor (Markdown or HTML)
• Collaboration: Share projects with editors/viewers, threaded annotations with @mentions
• Notifications: In-app bell + optional email alerts for OCR completion, shares, replies, mentions
• Version history: Every save creates a snapshot with visual diff and one-click revert
• Split & merge: Reorganize documents by splitting or merging texts
• Search: Full-text search across all owned and shared projects
• Math support: TeX input rendered with KaTeX, converted to MathML at Manifold export
• Metadata: Dublin Core fields for scholarly cataloging
• Export: ZIP archives with structured content and images for Manifold import

Project Structure

server/           Express API (port 3000)
  routes/         Route handlers (auth, projects, texts, ocr, llm, export, shares, annotations, notifications)
  services/       AI services (bedrock, openrouter), email, storage, cleanup
  middleware/     Auth, CSRF, rate limits, access control, upload validation
client/           React + Vite SPA
  src/pages/      Page components (Dashboard, ProjectView, TextDetail, etc.)
  src/components/ Shared components (Header, Footer, SharePanel, AnnotationSidebar, etc.)
  src/api/        API client
  src/lib/        PDF conversion pipeline
data/             Runtime data (SQLite DBs, uploaded files) — gitignored

License

This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

You are free to share and adapt this software for non-commercial purposes, provided you give appropriate credit to the CUNY AI Lab and distribute any derivative works under the same license.

See LICENSE for the full license text.

Contact

ailab@gc.cuny.edu