Legal House - AI-Powered Legal Document Management System

Legal House is an intelligent legal document review system that automates case analysis, generates AI-powered summaries of legal documents, and provides quick insights for law firms, courts, and legal professionals. The system leverages advanced natural language processing to streamline the legal workflow.

🚀 Features

Core Features

User Management
- Create and manage user profiles with contact details
- Store user information (name, email, phone, address)
- Link users to their respective cases
Case Management
- Create new legal cases with detailed information
- Upload legal documents (PDF format) for each case
- Track case creation and resolution timestamps
- Monitor case status (resolved/unresolved)
Document Upload & Storage
- Secure PDF file upload functionality
- File validation and size restrictions (10MB limit)
- Organized file storage system with unique filenames
- Direct access to uploaded documents via links
AI-Powered Document Summarization
- Automated text extraction from PDF documents
- Intelligent case summarization using local AI model (Ollama)
- Point-wise structured summaries for easy comprehension
- Real-time streaming response from AI model
Case Status Tracking
- Differentiate between resolved and unresolved cases
- Automatic status updates upon case resolution
- Timestamp tracking for creation and resolution
- Dashboard with case statistics
View & Filter Cases
- View all unresolved cases requiring attention
- Access resolved cases with their summaries
- Quick access to case documents
- Detailed case information display
Dashboard Analytics
- Active cases count
- Resolved cases count
- Total documents processed
- Recent cases overview

🛠 Tech Stack

Frontend

React.js (v19.0.0) - Modern JavaScript library for building user interfaces
Vite (v6.2.0) - Fast build tool and development server
React Router DOM (v7.4.1) - Client-side routing
Tailwind CSS (v3.4.1) - Utility-first CSS framework
ShadCN UI - Reusable component library built on Radix UI
Lucide React (v0.485.0) - Beautiful icon library
Axios (v1.8.4) - HTTP client for API requests
Sonner (v2.0.2) - Toast notifications
Marked (v15.0.7) - Markdown parser (for formatted summaries)

Backend

Node.js - JavaScript runtime environment
Express.js (v4.21.2) - Web application framework
PostgreSQL - Relational database management system
node-postgres (pg v8.14.1) - PostgreSQL client for Node.js
Multer (v1.4.5-lts.2) - Middleware for handling file uploads
pdf-parse (v1.1.1) - PDF text extraction library
node-fetch (v3.3.2) - Fetch API for Node.js
CORS (v2.8.5) - Cross-Origin Resource Sharing middleware
dotenv (v16.4.7) - Environment variable management

AI/ML Model

Ollama - Local AI model server
Deepseek R1 (1.56B) - Pretrained language model for text summarization
- Model Type: Generative text model
- Running locally on port 11434
- Streaming response capability
- Optimized for document summarization tasks

Database

PostgreSQL - Production-grade relational database
- Connection pooling for efficient resource management
- Foreign key constraints for data integrity
- Timestamp tracking for audit trails

🏗 System Architecture

┌─────────────────┐
│   React.js UI   │
│   (Frontend)    │
└────────┬────────┘
         │ HTTP Requests (Axios)
         ↓
┌─────────────────┐
│  Express.js API │
│   (Backend)     │
└────┬───────┬────┘
     │       │
     ↓       ↓
┌─────────┐ ┌──────────────┐
│PostgreSQL│ │Ollama AI Model│
│ Database │ │  (Local)      │
└──────────┘ └──────────────┘

Request Flow:

User Interaction → React frontend receives user input
API Request → Axios sends HTTP request to Express backend
Data Processing → Express controllers handle business logic
Database Operations → PostgreSQL stores/retrieves data
AI Processing → Ollama model generates summaries
Response → Data sent back through the chain to frontend

🗄 Database Schema

1. Users Table

Stores information about legal professionals, clients, or case owners.

CREATE TABLE users (
    id SERIAL PRIMARY KEY,
    user_name VARCHAR(255) NOT NULL,
    phone_no VARCHAR(15) NOT NULL,
    email VARCHAR(255) NOT NULL,
    address TEXT NOT NULL
);

Fields:

id - Auto-incrementing primary key
user_name - Full name of the user
phone_no - Contact phone number
email - Email address
address - Physical address

2. Cases Table

Stores legal case information and links to users.

CREATE TABLE cases (
    id SERIAL PRIMARY KEY,
    user_id INT NOT NULL,
    case_name VARCHAR(255) NOT NULL,
    case_file_link VARCHAR(255) NOT NULL,
    case_resolve_file_link TEXT,
    creation_timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    resolved_timestamp TIMESTAMP,
    resolve_status BOOLEAN DEFAULT FALSE,
    FOREIGN KEY (user_id) REFERENCES users(id) ON DELETE CASCADE
);

Fields:

id - Auto-incrementing primary key
user_id - Foreign key linking to users table
case_name - Name/title of the legal case
case_file_link - Path to uploaded PDF document
case_resolve_file_link - AI-generated summary text
creation_timestamp - Automatic timestamp when case is created
resolved_timestamp - Timestamp when case is resolved
resolve_status - Boolean indicating if case is resolved (true/false)

Relationships:

One user can have multiple cases (One-to-Many)
CASCADE delete ensures cases are removed when user is deleted

📡 API Documentation

Base URL

http://localhost:5000/api/v1

Endpoints

1. Create User

POST /create/user

Creates a new user in the system.

Request Body:

{
  "user_name": "John Doe",
  "phone_no": "1234567890",
  "email": "john@example.com",
  "address": "123 Main Street, City"
}

Response:

{
  "message": "User created successfully",
  "data": [
    {
      "id": 1,
      "user_name": "John Doe",
      "phone_no": "1234567890",
      "email": "john@example.com",
      "address": "123 Main Street, City"
    }
  ]
}

2. Create Case

POST /create/case

Creates a new case and uploads associated PDF document.

Content-Type: multipart/form-data

Request Body:

user_id (number) - ID of the user
case_name (string) - Name of the case
case_file_link (file) - PDF file (max 10MB)

Response:

{
  "message": "Case created successfully",
  "data": {
    "id": 1,
    "user_id": 1,
    "case_name": "Theft Case",
    "case_file_link": "/uploads/1234567890-document.pdf",
    "resolve_status": false,
    "creation_timestamp": "2024-01-06T10:30:00.000Z"
  }
}

3. Upload File

POST /create/upload

Uploads a PDF file to the server.

Content-Type: multipart/form-data

Request Body:

case_file_link (file) - PDF file

Response:

{
  "message": "File uploaded successfully",
  "file": "1234567890-document.pdf"
}

4. View Unresolved Cases

GET /view/view

Retrieves all cases that haven't been resolved yet.

Response:

{
  "message": "Cases retrieved successfully",
  "data": [
    {
      "case_id": 1,
      "case_name": "Theft Case",
      "case_file_link": "/uploads/1234567890-document.pdf",
      "creation_timestamp": "2024-01-06T10:30:00.000Z",
      "resolved_timestamp": null,
      "resolve_status": false,
      "user_id": 1,
      "user_name": "John Doe",
      "phone_no": "1234567890",
      "email": "john@example.com",
      "address": "123 Main Street, City"
    }
  ]
}

5. View Resolved Cases

GET /view/view-resolve

Retrieves all cases that have been resolved.

Response:

{
  "message": "Cases retrieved successfully",
  "data": [
    {
      "case_id": 2,
      "case_name": "Property Dispute",
      "case_file_link": "/uploads/9876543210-document.pdf",
      "creation_timestamp": "2024-01-05T09:00:00.000Z",
      "resolved_timestamp": "2024-01-06T11:00:00.000Z",
      "resolve_status": true,
      "user_id": 1,
      "user_name": "John Doe",
      "phone_no": "1234567890",
      "email": "john@example.com",
      "address": "123 Main Street, City"
    }
  ]
}

6. Get Case Summary

GET /view/get-summary?id={case_id}

Retrieves the AI-generated summary for a specific case.

Query Parameters:

id (number) - Case ID

Response:

{
  "success": true,
  "data": {
    "id": 2,
    "user_id": 1,
    "case_name": "Property Dispute",
    "case_file_link": "/uploads/9876543210-document.pdf",
    "case_resolve_file_link": "Summary text here...",
    "creation_timestamp": "2024-01-05T09:00:00.000Z",
    "resolved_timestamp": "2024-01-06T11:00:00.000Z",
    "resolve_status": true
  }
}

7. Summarize Case

POST /summarize

Processes a case document and generates an AI-powered summary.

Request Body:

{
  "pdfUrl": "/uploads/1234567890-document.pdf",
  "caseId": 1
}

Response:

{
  "success": true,
  "output": "**Case Summary:**\n1. Key point one...\n2. Key point two...\n\n**Conclusion:** Overall summary..."
}

Processing Steps:

Fetches PDF from the provided URL
Extracts text content from PDF
Sends text to Ollama AI model for summarization
Streams response from AI model
Updates database with summary and resolution status
Returns formatted summary to client

🔧 Installation & Setup

Prerequisites

Node.js (v16 or higher)
PostgreSQL (v12 or higher)
Ollama (for AI model)
npm or yarn package manager

1. Clone the Repository

git clone https://github.com/laatu08/SEMICOLON.git
cd SEMICOLON

2. Database Setup

# Install PostgreSQL (if not already installed)
# For Ubuntu/Debian:
sudo apt-get install postgresql postgresql-contrib

# For macOS:
brew install postgresql

# Start PostgreSQL service
sudo service postgresql start  # Linux
brew services start postgresql # macOS

# Create database
psql -U postgres
CREATE DATABASE legal_house;
\q

3. Backend Setup

cd backend

# Install dependencies
npm install

# Create .env file
touch .env

# Add the following to .env file:
DATABASE_URL=postgresql://username:password@localhost:5432/legal_house
PORT=5000
MODEL_NAME=deepseek-r1:1.5b

# Create uploads directory
mkdir uploads

# Start the server
npx nodemon index.js
# OR
npm start

The backend server will start on http://localhost:5000

4. Frontend Setup

# Open new terminal
cd admin

# Install dependencies
npm install

# Start the development server
npm run dev

The frontend application will start on http://localhost:5173

5. Ollama AI Model Setup

# Install Ollama
# For Linux:
curl -fsSL https://ollama.com/install.sh | sh

# For macOS:
brew install ollama

# For Windows: Download from https://ollama.com/download

# Start Ollama service
ollama serve

# Pull the Deepseek R1 model (in a new terminal)
ollama pull deepseek-r1:1.5b

# Verify model is running
curl http://localhost:11434/api/generate -d '{
  "model": "deepseek-r1:1.5b",
  "prompt": "Hello"
}'

🔄 Application Flow

Complete User Journey

1. Home Dashboard

User lands on homepage
↓
Sees statistics:
- Active cases count
- Resolved cases count
- Total documents
- Recent cases table

2. Create New Case Flow

User clicks "New Case" button
↓
Redirected to /create-case page
↓
User fills form:
- Full Name
- Email
- Phone Number
- Address
- Case Name
- PDF File Upload
↓
Form validation
↓
Submit form → API POST /api/v1/create/user
↓
User created in database
↓
API POST /api/v1/create/case (with user_id)
↓
File uploaded to /uploads directory
↓
Case created in database (resolve_status = false)
↓
Success notification shown

3. View & Summarize Cases Flow

User navigates to /view-case
↓
API GET /api/v1/view/view (fetches unresolved cases)
↓
Table displays all unresolved cases
↓
User clicks "Summarize" button
↓
Modal opens showing case details
↓
API POST /api/v1/summarize triggered
  ├─ Backend fetches PDF from server
  ├─ Extracts text using pdf-parse
  ├─ Sends text to Ollama AI model
  ├─ AI generates structured summary
  └─ Streams response back
↓
API updates database:
  ├─ Sets case_resolve_file_link = summary
  ├─ Sets resolved_timestamp = current time
  └─ Sets resolve_status = true
↓
Summary displayed in modal textarea
↓
User closes modal
↓
Case list refreshes (case removed from unresolved list)

4. View Resolved Cases Flow

User navigates to /resolve-case
↓
API GET /api/v1/view/view-resolve (fetches resolved cases)
↓
Table displays all resolved cases
↓
User clicks "View Summary" button
↓
API GET /api/v1/view/get-summary?id={case_id}
↓
Modal opens with pre-generated summary
↓
User can read the summary

Technical Flow Diagram

Frontend (React)
    │
    ├─→ Create Case
    │     ├─ Form Input
    │     ├─ File Upload
    │     └─ API Call → Backend → Database
    │
    ├─→ View Cases
    │     ├─ Fetch Data
    │     └─ API Call → Backend → Database Query
    │
    ├─→ Summarize
    │     ├─ Trigger AI
    │     ├─ API Call → Backend
    │     │     ├─ Fetch PDF
    │     │     ├─ Extract Text
    │     │     └─ Call Ollama
    │     └─ Update Database
    │
    └─→ View Summary
          └─ Fetch Summary → Backend → Database

🤖 AI Model Configuration

Ollama Integration

Model Details:

Name: Deepseek R1
Version: 1.5B parameters
Type: Generative Pre-trained Transformer
Purpose: Legal document summarization
Deployment: Local inference server

Configuration:

// API Endpoint
const OLLAMA_URL = "http://localhost:11434/api/generate";

// Model Parameters
{
  "model": "deepseek-r1:1.5b",
  "prompt": "<system_prompt> + <extracted_text>",
  "stream": true
}

Prompt Engineering: The system uses a carefully crafted prompt for optimal summarization:

You are an advanced AI model specialized in summarizing long texts.
Your task is to generate a structured, point-wise summary of the
given text, capturing key details concisely. Follow these guidelines:

1. Extract the most important points from the text in a clear and
   concise manner.
2. Present the summary in bullet points, avoiding unnecessary details.
3. Maintain clarity and coherence while ensuring completeness.
4. After listing the key points, provide a short conclusion
   summarizing the overall essence of the text.
5. Keep the response brief, to the point, and easy to read.

Here is the text to summarize:
[DOCUMENT_TEXT]

Response Processing:

The model returns a streaming response
Responses are concatenated into a full summary
<think> tags are removed from the output
Final summary is stored in the database

Model Performance:

Average processing time: 30-60 seconds per document
Suitable for documents up to 50 pages
Generates structured, point-wise summaries
Includes conclusion section

📁 Project Structure

SEMICOLON/
│
├── admin/                          # Frontend React Application
│   ├── public/
│   │   └── vite.svg
│   ├── src/
│   │   ├── assets/
│   │   │   └── react.svg
│   │   ├── components/
│   │   │   └── ui/                # ShadCN UI Components
│   │   │       ├── dialog.jsx     # Modal dialog component
│   │   │       └── navigation-menu.jsx
│   │   ├── lib/
│   │   │   └── utils.js          # Utility functions
│   │   ├── page/
│   │   │   ├── CreateCase.jsx    # Create new case page
│   │   │   ├── Footer.jsx        # Footer component
│   │   │   ├── Home.jsx          # Dashboard home page
│   │   │   ├── Navbar.jsx        # Navigation bar
│   │   │   ├── ResolveCase.jsx   # View resolved cases
│   │   │   └── ViewCase.jsx      # View & summarize cases
│   │   ├── App.css
│   │   ├── App.jsx               # Main app component with routing
│   │   ├── index.css             # Global styles
│   │   └── main.jsx              # Entry point
│   ├── .gitignore
│   ├── components.json           # ShadCN configuration
│   ├── eslint.config.js
│   ├── index.html
│   ├── jsconfig.json
│   ├── package.json
│   ├── postcss.config.js
│   ├── README.md
│   ├── tailwind.config.js
│   └── vite.config.js
│
├── backend/                        # Backend Node.js Application
│   ├── controllers/               # Business logic handlers
│   │   ├── case.controller.js    # Case creation logic
│   │   ├── fileUpload.controller.js  # File upload handling
│   │   ├── resolveView.controller.js # Resolved cases view
│   │   ├── summarize.controller.js   # AI summarization logic
│   │   ├── user.controller.js    # User management
│   │   └── view.controller.js    # Unresolved cases view
│   ├── models/                    # Database models
│   │   ├── case.model.js         # Cases table schema
│   │   └── user.model.js         # Users table schema
│   ├── routes/                    # API route definitions
│   │   ├── create.route.js       # Create user/case routes
│   │   ├── summerize.route.js    # Summarization route
│   │   └── view.route.js         # View cases routes
│   ├── uploads/                   # Uploaded PDF files storage
│   ├── .env                       # Environment variables
│   ├── .gitignore
│   ├── db.js                      # Database connection
│   ├── index.js                   # Server entry point
│   └── package.json
│
├── .gitignore
└── README.md                      # This file

Key Files Explanation

Frontend:

App.jsx - Main routing configuration
Home.jsx - Dashboard with statistics and recent cases
CreateCase.jsx - Form for creating new cases
ViewCase.jsx - Table of unresolved cases with summarize button
ResolveCase.jsx - Table of resolved cases with summaries
dialog.jsx - Reusable modal component for displaying summaries

Backend:

index.js - Express server setup and middleware configuration
db.js - PostgreSQL connection pool
case.controller.js - Handles case creation and file uploads
summarize.controller.js - PDF extraction and AI summarization
user.controller.js - User CRUD operations
*.model.js - Database table creation scripts

🔐 Environment Variables

Backend (.env)

# Database Configuration
DATABASE_URL=postgresql://username:password@localhost:5432/legal_house

# Server Configuration
PORT=5000

# AI Model Configuration
MODEL_NAME=deepseek-r1:1.5b

# Optional: Node Environment
NODE_ENV=development

Frontend (typically in .env or .env.local)

# API Base URL
VITE_API_URL=http://localhost:5000

Security Note: Never commit .env files to version control. Add them to .gitignore.

📖 Usage Guide

For Legal Professionals

Creating a Case

Navigate to the application homepage
Click on "New Case" button or go to "Create Case" in navigation
Fill in the client/case information:
- Enter full name of the client
- Provide email address
- Add phone number
- Enter physical address
- Specify case name/title
- Upload the legal document (PDF format, max 10MB)
Click "Create Case" button
Wait for success notification

Analyzing a Case

Go to "View Case" from navigation menu
Browse the list of unresolved cases
Click "View File" to review the original document
Click "Summarize" button for the desired case
Wait for AI processing (30-60 seconds)
Review the generated summary in the modal
The case is automatically marked as resolved

Reviewing Past Cases

Navigate to "Resolve Case" page
View all previously resolved cases
Click "View Summary" to see the AI-generated analysis
Review case details and summary

🤝 Contributors

Partha Borah - Lead Developer
Abishesk Prasad - Backend Developer
Nitul Das - Frontend Developer
Sourav Kumar Barman - Database & AI Integration

💡 Future Enhancements

Planned Features

Enhanced AI Capabilities
- Improve model efficiency for faster processing
- Support for larger documents (100+ pages)
- Multi-language document support
- Legal terminology recognition and highlighting
Advanced Search & Filtering
- Full-text search across cases
- Filter by date range, status, user
- Advanced query builder
- Export search results
Case Law Recommendations
- AI-powered similar case suggestions
- Legal precedent identification
- Citation extraction and linking
User Authentication
- Secure login system
- Role-based access control (Admin, Lawyer, Clerk)
- User activity logs
- Session management
Document Management
- Support for multiple file formats (DOCX, TXT, etc.)
- Document version control
- Annotation and highlighting features
- Collaborative editing
Reporting & Analytics
- Case resolution time analytics
- User activity reports
- PDF export of summaries
- Custom report generation
Notifications
- Email notifications for case updates
- Real-time alerts
- Reminder system for pending cases
Mobile Application
- React Native mobile app
- Offline access to summaries
- Push notifications
Integration Capabilities
- Calendar integration
- Email client integration
- Court management system APIs
- Cloud storage integration (Google Drive, Dropbox)
Security Enhancements
- End-to-end encryption for documents
- Two-factor authentication
- Audit trails
- GDPR compliance features

📜 License

This project is licensed under the MIT License.

MIT License

Copyright (c) 2024 SEMICOLON Team

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

📩 Contact

For queries, suggestions, or collaboration opportunities:

GitHub Repository: https://github.com/laatu08/SEMICOLON
Email: partha2002borah@gmail.com
Phone: +91 88128-63574
Location: JEC, Garmur, 785007

🙏 Acknowledgments

OpenAI and Ollama teams for AI model infrastructure
PostgreSQL community for robust database system
React and Vite teams for modern frontend tooling
ShadCN for beautiful UI components
All contributors and testers

📊 System Requirements

Minimum Requirements

CPU: 2 cores
RAM: 4 GB
Storage: 10 GB free space
OS: Windows 10, macOS 10.15+, Ubuntu 20.04+

Recommended Requirements

CPU: 4 cores or more
RAM: 8 GB or more
Storage: 20 GB SSD
GPU: Optional, for faster AI processing
OS: Latest stable versions

🐛 Troubleshooting

Common Issues

Issue 1: Database Connection Failed

Solution:
- Check PostgreSQL is running: sudo service postgresql status
- Verify DATABASE_URL in .env file
- Ensure database exists: psql -U postgres -c "\l"

Issue 2: Ollama Model Not Found

Solution:
- Check Ollama is running: curl http://localhost:11434
- Pull model: ollama pull deepseek-r1:1.5b
- Verify model name in .env matches exactly

Issue 3: File Upload Fails

Solution:
- Check uploads directory exists
- Verify file size is under 10MB
- Ensure file is PDF format
- Check disk space available

Issue 4: CORS Errors

Solution:
- Verify CORS is enabled in backend
- Check frontend is running on correct port
- Clear browser cache

"Empowering legal professionals with AI-driven automation!"

Last Updated: January 6, 2026 Version: 1.0.0 Maintained by: SEMICOLON Team

Name		Name	Last commit message	Last commit date
Latest commit History 35 Commits
admin		admin
backend		backend
.gitignore		.gitignore
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation

Legal House - AI-Powered Legal Document Management System

📋 Table of Contents

🚀 Features

Core Features

🛠 Tech Stack

Frontend

Backend

AI/ML Model

Database

🏗 System Architecture

Request Flow:

🗄 Database Schema

1. Users Table

2. Cases Table

📡 API Documentation

Base URL

Endpoints

1. Create User

2. Create Case

3. Upload File

4. View Unresolved Cases

5. View Resolved Cases

6. Get Case Summary

7. Summarize Case

🔧 Installation & Setup

Prerequisites

1. Clone the Repository

2. Database Setup

3. Backend Setup

4. Frontend Setup

5. Ollama AI Model Setup

🔄 Application Flow

Complete User Journey

1. Home Dashboard

2. Create New Case Flow

3. View & Summarize Cases Flow

4. View Resolved Cases Flow

Technical Flow Diagram

🤖 AI Model Configuration

Ollama Integration

📁 Project Structure

Key Files Explanation

🔐 Environment Variables

Backend (.env)

Frontend (typically in .env or .env.local)

📖 Usage Guide

For Legal Professionals

Creating a Case

Analyzing a Case

Reviewing Past Cases

🤝 Contributors

💡 Future Enhancements

Planned Features

📜 License

📩 Contact

🙏 Acknowledgments

📊 System Requirements

Minimum Requirements

Recommended Requirements

🐛 Troubleshooting

Common Issues

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Packages