Inspiration
I noticed that online shopping can be confusing and time-consuming. Users often need to visit multiple websites, compare product specifications, check prices, analyze product reliability, and make purchasing decisions manually. I wanted to solve this problem by building an AI-powered shopping assistant that simplifies the entire shopping experience and helps users make smarter and more informed buying decisions.
What it does
ShopPilot AI is an AI-powered autonomous shopping assistant that uses Google Gemini to understand natural-language shopping requests and assist users throughout their shopping journey.
Its key features include:
AI Shopping Agent: Understands user requirements and finds suitable products through natural-language queries. Smart Product Discovery: Recommends products based on category, specifications, and budget. Product Comparison: Compares products side by side, including processor, RAM, GPU, display, price, pros, and cons. Purchase Risk Analyzer: Evaluates available product information and identifies potential purchasing risks. Shopping Cart: Allows users to add products and manage their selections. PayPal Sandbox Integration: Demonstrates a checkout workflow with explicit user approval before payment capture. Order Management: Stores order and transaction information in MongoDB Atlas. Budget Protection: Respects user-defined budget limits and avoids silently recommending out-of-budget alternatives.
How we built it
I built ShopPilot AI using Google AI Studio and a full-stack development approach. I used Google Gemini 3.8 Flash with function calling to understand shopping queries and connect AI responses with backend tools. The frontend is built using React and TypeScript, while Node.js handles backend operations and API integration. I integrated MongoDB Atlas for storing product information, users, carts, orders, transactions, and shopping history. I also integrated PayPal Sandbox to demonstrate the checkout process, including payment approval, order creation, transaction recording, and duplicate payment prevention.
Challenges we ran into
During development, I faced several technical challenges: Integrating Gemini function calling with backend tools. Connecting and configuring MongoDB Atlas authentication. Maintaining accurate product information during search and comparison. Implementing budget restrictions in product discovery. Designing a safe payment workflow with explicit user approval. Preventing duplicate payment captures. Handling unavailable products and API errors properly. I addressed these challenges through debugging, testing, and improving the application's backend workflows.
Accomplishments that we're proud of
Successfully developed a functional AI-powered shopping assistant. Integrated Google Gemini function calling. Implemented intelligent product discovery and comparison. Established MongoDB Atlas connectivity with application collections. Integrated PayPal Sandbox checkout. Successfully tested all 8 end-to-end checkout scenarios, including duplicate payment prevention. Developed a modern and user-friendly shopping interface.
What we learned
While developing ShopPilot AI, I gained practical experience in AI agent development, Gemini function calling, React, TypeScript, Node.js, MongoDB Atlas, REST APIs, and payment integration. Most importantly, I learned how to connect AI models with real application operations and build an AI-powered application that goes beyond a basic chatbot.
What's next for ShopPilot AI
I plan to improve ShopPilot AI further by adding: Real-time product prices through e-commerce APIs. Price tracking and price-drop notifications. Personalized shopping recommendations. Advanced seller verification and product risk analysis. Improved order tracking. Additional payment gateway integrations.
Built With
- express.js
- google-gemini
- javascript
- mongodb-atlas
- node.js
- paypal-sandbox
- python
- react
- rest
- typescript
- vite