logo
ai negotiation
deal
items
landing

SwapU: The Autonomous Campus Marketplace

Inspiration

The spark for SwapU came from two places: the "campus graveyard" of unused student gear and our own personal passion as trading card collectors. We grew up mastering the art of the trade—evaluating rarity, negotiating value, and finding that perfect swap to complete a set. We realized that university life is surprisingly similar; students are constantly sitting on items/assets like iClickers, lab equipment, or textbooks that they no longer need, while others are scouring forums and discord channels to find those exact items.

Traditional marketplaces are too slow for the fast-paced student schedule. We wanted to take the thrill and fairness of card trading and apply it to campus essentials, removing the awkwardness of manual haggling and replacing it with an intelligent, autonomous system.

What it does

SwapU is an autonomous exchange platform designed for university students to trade lab gear, tech, and supplies without the friction of manual negotiation.

Autonomous Agents (Fetch.ai): Every user is represented by an agent that proactively proposes, evaluates, and negotiates trade combinations based on the user's inventory.
Browser-Informed Pricing(BrowserUse): To ensure a fair trade, our agents navigate the web and determine real-time market prices for items like iClickers or specific lab kits.
Real-time Coordination: Once the agents reach a "handshake" agreement, students use the integrated chat to finalize the physical hand-off on campus.
Persona Switching: A guest account selector on the Profile page allows for instant switching between different student personas to witness both sides of an autonomous negotiation in real-time.

How we built it

We built a unified stack that bridges autonomous agent logic with a modern web interface:

Agent Infrastructure: Fetch.ai powers the negotiation logic, allowing agents to act as independent economic proxies for students.
Backend: A FastAPI (Python 3.11+) server manages core domains including negotiation, deals, and items.
Frontend: Built with React 18 and Vite, utilizing Tailwind CSS v4 and shadcn/ui for a clean, highly responsive UI.
Data & Storage: Supabase (PostgreSQL) handles our relational database needs, while AWS S3 stores images of the items up for trade.

Challenges we ran into

Autonomous Agent Scaling: We hit a significant bottleneck with Browser-Use, which restricts the system to 10 concurrent sessions. We had to implement custom logic to populate data slowly and manage session queues to prevent crashes.
Fetch.ai Stability: Maintaining a consistent runtime for the Fetch.ai API proved difficult. Ensuring that the autonomous agents remained active and responsive to trade proposals throughout the negotiation lifecycle required constant monitoring.
Database Schema Constraints: We initially planned to store item lists as variables within our schemas, but discovered that PostgreSQL does not support lists in that format. This forced us to pivot our architecture and create additional relational tables to handle item management and trade combinations properly.
Rapid Full-Stack Learning Curve: As a relatively inexperienced team, we had to learn the intricacies of building a production-grade FastAPI backend on the fly, including mastering API endpoints and controllers while actively developing.

What we learned

This project taught us how to build Autonomous Economic Agents that represent true user intent. We moved beyond simple automation to explore how agents can navigate the web to make informed decisions. We also mastered a unified project structure where one Vite configuration manages a complex React frontend communicating with a high-performance Python backend.

What's next for SwapU

We plan to implement Multi-item Circular Trades, allowing agents to find three-way exchanges (e.g., Student A gives a lab coat to B, B gives a calculator to C, C gives a textbook to A) to maximize campus utility. We also aim to integrate university SSO to ensure SwapU remains a safe, verified environment for student-to-student exchange.