Inspiration
Online shopping often involves switching between multiple websites, comparing prices, managing purchases, and coordinating shared expenses. We wanted to simplify this experience by combining AI-powered shopping assistance with the idea of collaborative spending, inspired by UPI Circle. Our vision was to make shopping more intelligent while ensuring users retain control over their money.
What it does
BuyPal is an AI-powered shopping assistant with a vision for collaborative spending.
- Helps users discover products and compare prices across merchants.
- Uses AI to understand shopping intent and orchestrate product discovery.
- Explores shared spending with configurable permissions, spending limits, and expiry controls.
- Integrates with PayPal to explore secure payment and checkout workflows.
- Keeps users in control through approval-required checkout by default, with optional bounded autonomous purchasing.
How we built it
We designed BuyPal as a modular application with separate components for the frontend, backend, AI agent, and payment processing.
- Frontend: React for the user interface and shopping experience.
- Backend: Node.js for application APIs and business logic.
- AI Agent: Handles product discovery, comparison, and purchase-intent orchestration.
- Payment Service: Integrates with PayPal through official payment and authorization flows.
- Architecture: Modular services with server-side permission checks and a focus on security, extensibility, and maintainability.
Challenges we ran into
One of our biggest challenges was designing an autonomous shopping experience without compromising payment security or user control.
- Separating AI-driven product discovery from actual payment authorization.
- Designing delegated spending workflows while accounting for PayPal API limitations and eligibility.
- Ensuring spending limits and permissions are enforced server-side rather than trusted to the AI agent.
- Coordinating multiple independently developed components into one consistent application.
- Balancing automation with transparency and explicit user consent.
Accomplishments that we're proud of
- Designed a unified vision combining AI-assisted shopping and collaborative spending.
- Established a modular architecture that separates AI orchestration from payment authorization.
- Prioritized user consent, spending limits, and revocable permissions as core design principles.
- Created a foundation for exploring agentic commerce through PayPal integration.
- Established a collaborative development workflow that allows different team members to work independently.
What we learned
We learned that building an autonomous shopping agent involves much more than connecting an LLM to a browser. Reliable product discovery, payment authorization, security boundaries, and user consent must work together.
We also learned the importance of validating payment-provider capabilities against official APIs instead of assuming that familiar payment features are universally available.
Most importantly, autonomy should not mean giving up control. AI should handle the complexity while users retain authority over financial decisions.
What's next for BuyPal
- Expand live product discovery and cross-merchant price comparison.
- Improve browser automation and shopping workflow reliability.
- Validate delegated spending capabilities against PayPal's official APIs and sandbox.
- Implement configurable spending policies, expiry, and revocation.
- Strengthen purchase auditing and server-side authorization.
- Explore bounded autonomous checkout with explicit opt-in and spending constraints.
- Move toward a seamless, secure agentic commerce experience.
