Inspiration
Online shopping can be confusing and time-consuming. Customers often visit multiple websites to compare prices, specifications, reviews, and product features. We wanted to solve this problem by creating an AI-powered shopping assistant that makes purchasing smarter, faster, and more convenient.
What it does
AgentCart AI is an autonomous AI shopping agent that understands customer needs and helps them make better purchasing decisions.
- Finds products based on natural-language requests.
- Compares prices, features, and specifications.
- Recommends products based on budget and preferences.
- Optimizes spending by suggesting the best value for money.
- Provides a shopping cart, wishlist, and order history.
- Integrates PayPal for secure checkout with explicit customer approval.
How we built it
We designed AgentCart AI using a modern full-stack architecture:
- Frontend: React/Next.js, TypeScript, and Tailwind CSS.
- Backend: Node.js and Express APIs.
- AI: Google Gemini API for natural-language understanding, product recommendations, and intelligent comparisons.
- Database: PostgreSQL for users, products, carts, and orders.
- Payments: PayPal Orders API and Sandbox integration.
- Deployment: GitHub, Vercel, and Render.
The AI agent processes shopping requests, evaluates available product data, and generates structured recommendations. The customer remains in control and approves every purchase before payment.
Challenges we ran into
- Designing an AI agent that understands different shopping requirements.
- Generating reliable recommendations using actual product information.
- Integrating AI services with backend APIs and the database.
- Implementing secure PayPal payment workflows.
- Managing customer budgets while finding the best product value.
- Ensuring security, responsiveness, and a smooth user experience.
Accomplishments that we're proud of
- Creating an AI-powered concept that transforms traditional online shopping.
- Combining product discovery, comparison, and budget optimization in one platform.
- Designing an agentic shopping workflow with customer-controlled purchasing.
- Integrating PayPal as the secure payment solution.
- Building a scalable foundation that can support multiple product categories and merchants.
What we learned
We learned how agentic AI can simplify real-world shopping decisions, how to connect AI models with full-stack applications, and why secure payment verification and customer authorization are essential in AI-powered commerce.
What's next for AGENTCART AI
Our next goal is to develop a fully functional autonomous shopping experience with real-time merchant catalogs, more personalized recommendations, advanced price tracking, multi-store comparison, voice-based shopping, and expanded payment options. We aim to make AgentCart AI a trusted intelligent shopping companion for everyone.
Built With
- ai
- css
- express.js
- github
- google-gemini-api
- html
- javascript
- neon-database
- next.js
- node.js
- paypal
- paypal-sandbox
- postgresql
- prisma-orm
- react
- render
- rest-api
- tailwind-css
- typescript
- vercel