Veridox: AI-Powered Legal Contract Risk Analysis Platform

Veridox is an AI-powered web application focused on legal contract risk analysis. Users can upload contracts in PDF format; after text extraction via OCR, large language models are used to identify potential risks, compliance issues, and key clauses. It uses a React frontend + Spring Boot backend architecture, with enterprise-level security design, aiming to help individuals and small businesses reduce legal risks and improve contract review efficiency.

• Original Author/Maintainer: AnimeshJ004
• Source Platform: GitHub
• Original Link: https://github.com/AnimeshJ004/Veridox
• Online Demo: https://contract-risk-analyzer-theta.vercel.app/
• Release Date: May 24, 2026

Veridox is positioned as an AI-driven tool for legal contract analysis, using a decoupled frontend-backend architecture with a modern and enterprise-level tech stack.

Security Authentication System

Uses JWT token management based on HttpOnly Cookies, supporting email OTP verification, password reset, and Google OAuth2 third-party login.

AI-Powered Contract Analysis

After uploading a PDF, text is extracted via Tesseract OCR; Gemini/Groq large models are called to perform risk level assessment, jurisdiction compliance checks, and clause parsing.

Smart Dashboard

Supports contract list viewing, risk badge identification, status tracking, and historical data analysis.

Admin Console

Admins can monitor system usage (statistics on number of uploaded files, size, etc.).

Rate Limiting and Quota Management

Regular users have daily upload quota limits; the interface displays a real-time countdown.

Interactive AI Dialogue

Users can have real-time conversations with AI to ask legal questions and get clause summaries.

Frontend Tech Stack

React.js (Vite), React-Bootstrap & Tailwind, Axios, React-Router-DOM, PDF-lib.

Backend Tech Stack

Java 17, Spring Boot 3.x (including Web/Data MongoDB/Security/Mail/OAuth2 modules), Tesseract OCR, Gemini/Groq large model integration.

Database

MongoDB / MongoDB Atlas.

Security Design

• Stateless session management (JWT for authentication information transmission)
• XSS protection (token stored in HttpOnly Cookies)
• CSRF protection (configured via Spring Security)
• Auto-logout (session destroyed after 30 minutes of inactivity)

For individuals and small businesses, the value includes:

1. Reducing legal risks: Identify potential issues before signing
2. Improving review efficiency: AI completes preliminary analysis in seconds
3. Popularizing legal knowledge: Interactive dialogue aids learning
4. Cost control: Replace expensive full-time legal staff

1. Organic combination of OCR and LLM: Tesseract text conversion + LLM semantic analysis
2. Enterprise-level security practices: Comprehensive protection mechanisms ensure sensitive document security
3. Modern decoupled frontend-backend: Optimized implementation using mainstream tech stack (React + Spring Boot)

Veridox is a valuable attempt of AI in the legal service field, with solid technology and comprehensive security, suitable for developers to learn. Future expansion directions:

• Support more document formats
• Integrate more regional legal databases
• Provide multi-language support