RoofProof

• 

  Landing Page

• 

  House Listing Page

• 

  Browse Homes Page

• 

  Anomaly scan

• 

  Verified Document

• 

  Application Page

Inspiration

Renting a property often relies on documents such as salary slips and Form 16, but forged or manipulated documents can make it difficult for landlords to verify tenants with confidence. We wanted to build a solution that makes rental verification safer, faster, and more trustworthy while respecting tenant privacy.

What it does

RoofProof is an AI-powered rental verification platform that helps landlords identify potentially fraudulent tenant documents and make safer rental decisions.

It uses OCR to extract information from documents, AI-based anomaly detection to identify suspicious inconsistencies, and Midnight blockchain as a privacy-preserving verification layer for sensitive verification data.

How we built it

We built RoofProof as a web-based platform connecting the frontend, backend, AI verification pipeline, and blockchain layer.

• OCR extracts structured information from uploaded documents.
• AI anomaly detection analyzes Form 16 and other financial documents for suspicious patterns or inconsistencies.
• Midnight provides a privacy-focused blockchain layer for verification without unnecessarily exposing sensitive tenant information.
• A web interface brings the entire verification workflow together for landlords and tenants.

Challenges we ran into

Our biggest challenge was balancing fraud detection, usability, and privacy. Financial documents contain sensitive information, so we had to carefully decide what should be processed, what should be verified, and what should never be exposed.

We also faced challenges with different document formats, OCR accuracy, identifying meaningful anomalies, and integrating multiple technologies into a single workflow.

Accomplishments that we're proud of

We are proud of building a working end-to-end prototype that brings AI-powered document verification and privacy-preserving blockchain technology together for a real-world rental problem.

We also implemented anomaly detection specifically to help identify potentially fake or manipulated Form 16 documents, rather than relying only on basic document extraction.

What we learned

RoofProof taught us how to combine AI, OCR, web development, and blockchain into a practical application.

We also learned that blockchain doesn't have to mean putting sensitive information on-chain. With Midnight, we explored how blockchain can instead act as a privacy-focused verification layer, which is especially important when dealing with financial and identity-related documents.

What's next for RoofProof

We want to expand RoofProof with stronger document fraud detection, more document types, improved verification accuracy, and additional privacy-preserving verification workflows.

Our long-term goal is to make RoofProof a trusted rental verification ecosystem where landlords can verify confidently while tenants retain control over their sensitive information.

Built With

• compact-0.2.0
• express.js
• git
• github
• javascript-(es6+)
• lucide-react-icons
• machine-learning-(form-16-ai-anomaly-detector)
• midnight-js-sdk
• midnight-network
• neon-cloud-postgresql
• node.js
• npm-workspaces
• postgresql
• react-18
• vanilla-css
• vite-5
• zero-knowledge-proofs-(zkps)