BuySmart: AI-Assisted E-Commerce Platform

Star History

Project Overview

BuySmart is a comprehensive e-commerce platform developed as a CSE327.2 Software Engineering project. The system integrates multiple AI agents to enhance user experience while maintaining transparency, security, and reliability through sound software engineering practices.

Built using agile development methodology over 6 weeks (3 sprints), the platform demonstrates disciplined AI integration where core business logic remains deterministic while AI agents provide decision support and user assistance.

Team: 3 members
Duration: 6 weeks (Feb 8 - March 21, 2026)
Methodology: Agile Scrum with 2-week sprints

Key Features

Core E-Commerce Functionality

• User authentication with email/password and federated login (Google, Facebook)
• Product catalog with search, filtering, and category management
• Shopping cart and checkout system
• Order management and tracking
• Seller dashboard for product management
• Buyer dashboard with order history

AI-Powered Enhancements

• Smart Product Recommendations: Personalized suggestions based on user preferences and browsing history
• Feedback Sentiment Analysis: Automated analysis of customer reviews for quality insights
• Refund Decision Support: AI-assisted refund eligibility evaluation with human-in-the-loop approval
• Customer Support Chatbot: Multi-agent orchestration for FAQ handling and issue escalation

Administration & Transparency

• Admin dashboard with user and refund management
• Comprehensive activity logging and audit trails
• AI action approval workflow for sensitive operations
• Real-time notifications and system monitoring

Tech Stack

Frontend

• Next.js 14
• React 18
• Tailwind CSS
• shadcn/ui

Backend & Database

• Supabase (Edge Functions, RLS policies)
• PostgreSQL (Supabase)
• Supabase Auth
• Supabase Storage & Realtime

AI & Integration

• LangChain (Python/TypeScript)
• TanStack Query
• React Context (state management)

Development & Testing

• ESLint, Prettier, Husky
• Jest/Vitest (frontend tests)
• Playwright (E2E tests)

Deployment

• Vercel (frontend)
• Supabase (backend + database)

Architecture Overview

System Flow

• Frontend communicates with Supabase via client APIs for authentication, data queries, and real-time updates
• Edge Functions handle complex business logic for orders, refunds, and AI integrations
• Row Level Security (RLS) policies enforce role-based access control for buyers, sellers, and admins
• Database triggers maintain comprehensive activity logs for audit and transparency

Data Layer

Core entities: users, products, orders, order_items, feedback, refund_requests, activity_logs, ai_agent_logs, chat_messages

AI Integration

• LangChain agents generate recommendations, sentiment analysis, refund assistance, and chatbot responses
• Human-in-the-loop workflow stores AI actions as pending until admin approval
• All AI decisions logged in ai_agent_logs for transparency and audit

Search & Notifications

• Full-text search using Supabase PostgreSQL capabilities
• In-app and email notifications via Supabase triggers
• Real-time dashboard updates using Supabase Realtime

Setup & Installation

Prerequisites

• Node.js (LTS version)
• Supabase account and project
• Vercel account (for deployment)

Environment Variables

Create a .env.local file with the following variables:

NEXT_PUBLIC_SUPABASE_URL=your_supabase_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key
SUPABASE_SERVICE_ROLE_KEY=your_service_role_key
SUPABASE_JWT_SECRET=your_jwt_secret
GOOGLE_OAUTH_CLIENT_ID=your_google_client_id
GOOGLE_OAUTH_CLIENT_SECRET=your_google_client_secret
FACEBOOK_OAUTH_CLIENT_ID=your_facebook_client_id
FACEBOOK_OAUTH_CLIENT_SECRET=your_facebook_client_secret
LANGCHAIN_API_KEY=your_langchain_api_key
AI_AGENT_MODEL=your_preferred_model

Local Development

1. Clone the repository
2. Install dependencies: npm install or pnpm install
3. Configure environment variables in .env.local
4. Start development server: npm run dev

Database Setup

1. Create a new Supabase project
2. Apply database schema for all core tables
3. Configure RLS policies for role-based access
4. Set up database triggers for activity logging
5. Seed initial demo data (optional)

Available Commands

Development

• npm run dev - Start development server
• pnpm dev - Alternative with pnpm

Build & Deploy

• npm run build - Production build
• pnpm build - Alternative with pnpm

Testing

• npm run test - Run unit tests (Jest/Vitest)
• npx playwright test - Run E2E tests

Code Quality

• npm run lint - Run ESLint
• npm run format - Run Prettier

Agile Development Process

The project follows a structured 3-sprint agile approach:

Sprint 1 (Weeks 1-2): Foundation setup, authentication system, product catalog, and basic AI recommendations
Sprint 2 (Weeks 3-4): Order management, feedback system, refund workflow, and admin dashboard
Sprint 3 (Weeks 5-6): Customer support chatbot, comprehensive testing, and production deployment

Each sprint delivers working software increments with continuous integration, testing, and stakeholder feedback. The team maintains a product backlog of 56 user stories across 9 epics, prioritizing high-value features for early delivery.

Demo / Screenshots

Screenshots and demo videos showcasing key user journeys:

Video Demo

Watch the full platform demo: BuySmart in Action

• User registration and authentication flow
• Product browsing and AI-powered recommendations
• Complete purchase workflow from cart to confirmation
• Seller dashboard and product management
• Admin interface with AI action approvals
• Customer support chatbot interactions