Sesame Exchange

A revolutionary cross-platform mobile application that leverages artificial intelligence to transform how students buy, sell, and exchange items on campus. Built with cutting-edge technology to make campus commerce seamless, sustainable, and smart.

Flutter
Firebase
Hugging Face AI
Cloudinary
Dart
95%+
AI Accuracy
7
Categories
<2s
Image Processing
Real-time
Messaging

The Challenge & Solution

Addressing a real problem with innovative technology

The Problem

$500+

Average unused items per student

40%

Campus waste from discarded items

Hours

Wasted on fragmented marketplaces

  • Students struggle to find trusted buyers/sellers on campus
  • Existing platforms lack proper categorization and search
  • Manual item listing is time-consuming and error-prone
  • No integrated communication for quick transactions
  • Environmental impact from throwing away reusable items

Our Solution

Sesame Exchange revolutionizes campus commerce with AI-powered automation and seamless user experience.

  • AI Image Recognition

    Hugging Face BLIP model automatically identifies and categorizes items with 95%+ accuracy

  • Smart Categorization

    Intelligent sorting across 7 categories: Electronics, Books, Furniture, Clothing, Sports, Kitchen, Other

  • Real-time Messaging

    Firebase-powered instant chat for seamless communication between buyers and sellers

  • Google OAuth

    Secure authentication with university email verification

  • Cloudinary CDN

    Lightning-fast image delivery with optimized caching and compression

$200-500

Potential savings per semester per student

30%

Projected reduction in campus waste

Campus-wide

Scalable to thousands of students

Sustainable

Promoting circular economy on campus

Key Features

Powerful functionality meets intuitive design

AI-Powered Image Recognition

Leveraging Hugging Face's state-of-the-art BLIP (Bootstrapping Language-Image Pre-training) model, Sesame Exchange automatically analyzes uploaded images to identify items and suggest accurate categories. The AI processes images in under 2 seconds with over 95% accuracy.

  • Instant item identification
  • Natural language descriptions
  • Automatic category suggestion
  • Title and description generation
Hugging Face BLIP Model ML Integration
AI Image Recognition in Action
AI automatically categorizing a laptop listing

Intelligent Feed & Smart Filtering

Browse items effortlessly with our dynamic feed that updates in real-time. Advanced filtering system allows users to narrow down results by category, price range, condition, and more. The responsive grid layout ensures optimal viewing on any device.

  • Real-time feed updates
  • 7 category filters
  • Instant search functionality
  • Sort by date, price, relevance
Firestore Queries Real-time Listeners Flutter ListView
Main Feed Interface Category Filters
Dynamic feed with smart category filtering

Seamless Real-time Messaging

Built on Firebase Realtime Database, our messaging system enables instant communication between buyers and sellers. Messages are delivered instantaneously with read receipts, typing indicators, and conversation threading for a smooth chat experience.

  • Instantaneous message delivery
  • Read receipts & status updates
  • Secure, encrypted conversations
  • Organized conversation threads
Firebase Realtime DB Stream Controllers WebSocket
Real-time Messaging Interface
Real-time chat with instant message delivery

Comprehensive User Profiles

Each user has a detailed profile showcasing their active listings, transaction history, and reputation score. Google OAuth integration ensures authentic campus community members while maintaining privacy and security.

  • Verified student authentication
  • Personal listing management
  • Reputation & rating system
  • Privacy-focused design
Google OAuth 2.0 Firebase Auth Firestore Security
User Profile Interface
Comprehensive user profile with listing management

Interactive Comments System

Foster community engagement with our robust commenting system. Users can ask questions, share insights, and build trust before making a transaction. Threaded replies and real-time updates keep conversations organized and active.

  • Public questions and answers
  • Threaded comment replies
  • Timestamped updates
  • Moderation capabilities
Firestore Collections Real-time Updates Flutter Widgets
Comments Section
Engaging comments system for community interaction

Technical Architecture

Built with modern, scalable technologies

System Architecture

Frontend Layer

Flutter UI
State Management
Navigation

Service Layer

Firebase Auth
Hugging Face API
Cloudinary SDK

Backend Layer

Firestore DB
Realtime Database
Cloud Storage

Frontend - Flutter & Dart

Flutter Framework

Cross-platform mobile development with a single codebase for iOS and Android. Leveraging Material Design 3 for modern, responsive UI components.

Hot Reload Widget Trees Stateful Management Custom Animations

State Management

Provider pattern for efficient state management across the app, ensuring smooth UI updates and minimal rebuilds.

Provider ChangeNotifier StreamBuilder

Backend - Firebase Suite

Firebase Authentication

Secure Google OAuth 2.0 integration with university email verification. Handles user sessions, tokens, and refresh mechanisms automatically.

Google Sign-In JWT Tokens Auto Refresh

Cloud Firestore

NoSQL document database for storing listings, user profiles, and comments. Real-time synchronization with offline persistence.

Real-time Sync Offline Mode Complex Queries Security Rules

Realtime Database

WebSocket-based messaging system for instant chat functionality. Sub-millisecond latency for true real-time communication.

WebSocket Low Latency Presence System

AI/ML - Hugging Face

BLIP Image Captioning Model

State-of-the-art vision-language model that generates natural language descriptions of images. Trained on millions of image-text pairs for accurate item identification.

Vision Transformers Zero-shot Learning 95%+ Accuracy

Custom Integration

Built custom API wrapper for seamless integration with Flutter app. Handles image encoding, API requests, and response parsing efficiently.

REST API Base64 Encoding Error Handling Retry Logic

Media - Cloudinary CDN

Image Optimization

Automatic image compression, format conversion, and responsive delivery. Reduces bandwidth usage by up to 70% while maintaining visual quality.

Auto WebP Lazy Loading Compression Responsive Images

Global CDN

Distributed edge servers ensure fast image loading worldwide. Intelligent caching reduces latency and improves user experience.

Edge Caching Global Distribution Auto Invalidation

Interested in This Project?

The repository is private, but I'd love to share the code with recruiters and collaborators. Reach out to request access!