🛡️ LexGuard - Contract Risk Analyzer for SMEs

LexGuard is a lightweight, privacy-first contract analysis tool designed to help entrepreneurs and small businesses quickly identify potentially risky clauses in agreements.

It uses rule-based pattern matching to highlight clauses related to liability, exclusivity, termination, and more; guiding users to the most important parts of a contract for further review.

🚀 Features

• 🔍 Clause Detection : Identifies key legal clauses using regex-based rules

• ⚠️ Risk Classification : Assigns High, Medium, Low risk levels

• 💬 Contextual Warnings : Explains why a clause may be risky

• 📊 Visual Analysis : Displays risk distribution (chart view)

• 📁 File Upload Support : Analyze .txt and .pdf contracts

• 🔐 Privacy-First : Runs entirely in the browser (no data storage)

• 📥 JSON Export : Download structured analysis results

🧠 Approach

LexGuard combines data-driven insights with rule-based NLP techniques.

Contract clause patterns were derived from structured legal text analysis

These patterns were translated into regex-based detection rules

Rules are organized into modular JSON files:

• rules.json

• severity_mapping.json

• warning_messages.json

This design allows the system to be easily updated without modifying core logic.

⚙️ Key Technical Decision

Initially, TF-IDF was explored alongside rule-based detection.

However:

It introduced noise by flagging common words (e.g., “agreement”, “terms”, "use")

This led to inflated and less meaningful risk scores

👉 The system was refined to use pure regex-based pattern matching, improving precision and interpretability.

🧩 System Design

LexGuard is built with a modular architecture:

• Frontend: HTML, CSS, JavaScript

• Analysis Engine: Regex-based clause detection

• Data Layer: JSON-driven rules and mappings

• Deployment: Netlify (frontend hosting)

⚡ Development Workflow

This project was built using a vibe coding / no-code assisted workflow, combining:

• ChatGPT

• Anti Gravity AI

• GitHub

• Netlify

This enabled rapid prototyping, debugging, and iteration.

🚧 Challenges & Learnings

• Managing compute and model access limitations

• Adapting by switching between AI models and optimizing prompts

• Debugging deployment issues (e.g., backend limitations on Netlify)

• Transitioning to a fully client-side architecture

🔐 Privacy by Design

LexGuard does not store or transmit contract data.

• All analysis runs locally in the browser

• No database or backend storage

• Users maintain full control of their documents

📦 Output

The system generates a structured JSON report:

{ "overall_risk": "Medium", "risk_score": 45, "risk_counts": { "High": 2, "Medium": 5, "Low": 3 }, "flags": [...] }

This allows for further analysis or integration into other workflows.

🎯 Use Cases

• SME contract review

• Pre-legal review screening

• Legal awareness and education

• Data-driven contract analysis

⚠️ Disclaimer

LexGuard is an rule-based contract analysis tool designed to support contract review.

It does not provide legal advice and should not replace consultation with a qualified legal professional.

🎬 Demo Video

👉 Click the image above to watch the demo.

🌍 Live Demo

👉 https://lexguard20.netlify.app/

👩🏽‍💻 Author

Latifah Bashir 3MTT Fellow (AI/ML Track) Data Science & Analytics Enthusiast