🛍️ Snitch Fullstack

A full-stack MERN E-Commerce Application inspired by the Snitch clothing brand. The project provides secure authentication, product management, cloud image uploads, shopping cart functionality, and AI-powered semantic product search using Mistral AI Embeddings and Pinecone Vector Database.

---

🚀 Features

🔐 Authentication

• User Registration
• User Login (JWT Authentication)
• Secure HTTP-only Cookies
• Logout
• Get Current User
• Protected Routes
• Google OAuth (In Progress)

---

📦 Product Management

• Create Product
• Update Product
• Delete Product
• Get All Products
• Get Product by ID
• Get Products by Category
• Get Distinct Categories
• Cloud Image Upload using ImageKit
• Product Stock Management

---

🛒 Shopping Cart

• Add Product to Cart
• Get Logged-in User Cart
• Remove Product from Cart
• Clear Cart
• Quantity Management
• Stock Validation before Adding Products
• Protected Cart APIs

---

🤖 AI Features

• AI-powered Semantic Product Search
• Product Embedding using Mistral AI Embeddings
• Vector Storage using Pinecone
• Automatic Embedding Generation when Product is Created
• Automatic Embedding Update when Product is Updated
• Automatic Vector Deletion when Product is Deleted
• Semantic Similarity Search using Vector Embeddings

---

🔒 Security

• JWT Authentication
• HTTP-only Cookies
• Password Hashing using bcrypt
• Protected REST APIs
• CORS Configuration

---

🛠 Tech Stack

Frontend

• React
• Vite
• React Router
• Axios
• Tailwind CSS

Backend

• Node.js
• Express.js
• MongoDB
• Mongoose
• Multer
• ImageKit
• JWT
• Cookie Parser

AI Stack

• LangChain
• Mistral AI Embeddings
• Pinecone Vector Database

---

📂 Project Structure

snitch-fullstack/
│
├── Backend/
│   ├── src/
│   │
│   ├── controllers/
│   │   ├── auth.controller.js
│   │   ├── post.controller.js
│   │   └── cart.controller.js
│   │
│   ├── middleware/
│   │
│   ├── models/
│   │   ├── user.models.js
│   │   ├── post.models.js
│   │   └── cart.models.js
│   │
│   ├── routes/
│   │   ├── auth.routes.js
│   │   ├── post.routes.js
│   │   └── cart.routes.js
│   │
│   ├── config/
│   ├── app.js
│   ├── package.json
│   └── .env
│
├── Frontend/
│   ├── src/
│   ├── public/
│   ├── package.json
│   └── vite.config.js
│
└── README.md

---

📦 Installation

Clone Repository

git clone https://github.com/sandeep7348/snitch-fullstack.git

cd snitch-fullstack

---

Backend Setup

cd Backend

npm install

Create a .env file.

PORT=3000

MONGODB_URL=your_mongodb_connection_string

JWT_SECRET=your_jwt_secret

IMAGE_KIT_PUBLIC_KEY=your_public_key
IMAGE_KIT_PRIVATE_KEY=your_private_key
IMAGE_KIT_URL_ENDPOINT=your_url_endpoint

GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret

MISTRAL_API_KEY=your_mistral_api_key

PINECONE_API_KEY=your_pinecone_api_key
PINECONE_INDEX_NAME=your_pinecone_index_name

Run Backend

npm run dev

---

Frontend Setup

cd Frontend

npm install

npm run dev

---

REST API

Authentication

Method	Endpoint
POST	/api/auth/register
POST	/api/auth/login
GET	/api/auth/getMe
POST	/api/auth/logout

---

Products

Method	Endpoint
POST	/api/post
GET	/api/allpost
GET	/api/post/:id
PUT	/api/post/:postId
DELETE	/api/post/:postId
GET	/api/category/:category
GET	/api/categories
POST	/api/search

---

Shopping Cart

Method	Endpoint
POST	/api/cart/add
GET	/api/cart
DELETE	/api/cart/remove/:postId
DELETE	/api/cart/clear

---

AI Semantic Search

Every product is converted into a vector embedding using Mistral AI Embeddings.

The embedding is stored in Pinecone along with product metadata.

When a user searches for a product:

• The search query is converted into an embedding.
• Pinecone performs semantic similarity search.
• Matching product IDs are returned.
• Complete product details are fetched from MongoDB.
• Relevant products are returned to the user.

---

Product Creation Flow

Admin Creates Product
        │
        ▼
 Upload Image (ImageKit)
        │
        ▼
 Save Product (MongoDB)
        │
        ▼
 Generate Embedding (Mistral AI)
        │
        ▼
 Store Vector (Pinecone)

---

Shopping Cart Flow

User Login
      │
      ▼
Browse Products
      │
      ▼
Click Add to Cart
      │
      ▼
Authenticate User
      │
      ▼
Check Product Availability
      │
      ▼
Validate Stock
      │
      ▼
Create Cart / Update Existing Cart
      │
      ▼
Save Cart in MongoDB
      │
      ▼
Return Updated Cart

---

Semantic Search Flow

User Search Query
        │
        ▼
Generate Query Embedding
        │
        ▼
Pinecone Similarity Search
        │
        ▼
Retrieve Matching Product IDs
        │
        ▼
Fetch Products from MongoDB
        │
        ▼
Return Relevant Products

---

Example Search Request

{
  "query": "black oversized cotton t-shirt"
}

---

Example Search Queries

• oversized black t-shirt
• formal white shirt
• winter hoodie
• casual streetwear
• denim jeans
• cotton summer shirt
• cargo pants
• office wear
• premium men's clothing
• comfortable everyday clothes

---

Image Upload

Product images are uploaded to ImageKit.

Only the secure image URL is stored in MongoDB.

---

Authentication Flow

Register/Login
      │
      ▼
Generate JWT
      │
      ▼
Store HTTP-only Cookie
      │
      ▼
Access Protected APIs

---

Upcoming Features

• Order Management
• Stripe / Razorpay Integration
• Wishlist
• AI Shopping Assistant
• Personalized Product Recommendations
• Product Filtering & Sorting
• Pagination
• Product Reviews & Ratings
• Admin Dashboard
• User Profile Management
• Google OAuth Authentication

---

Author

Sandeep Choudhary

GitHub: https://github.com/sandeep7348

---

⭐ If you found this project useful, consider giving it a star on GitHub!