Prophet

prophet logo

Inspiration We realized that by the time a trend hits TikTok or Instagram, it’s usually too late to start selling merchandise for it. We wanted to solve the "inventory risk" problem by looking forward instead of backward. We asked: "What if we could use prediction markets (people betting money on outcomes) to tell us exactly what to sell before it happens?"

What it does Prophet acts as an autonomous e-commerce manager. It constantly watches Polymarket for high-probability cultural events (like "Taylor Swift announces new album" or "Chiefs win the Superbowl"). When it spots a likely event, it automatically designs a relevant t-shirt (using AI) and publishes the product directly onto the shopify page

How we built it We used LangGraph for using our agents to take predication market data and generate products from them and used Shopify API to release those products onto the store ready to sell.

Challenges we ran into Our biggest challenge was navigating the Shopify API and correctly setting up authorization for our application.

Accomplishments that we're proud of We are really proud of the "Multi-Agent" workflow. Seeing the agents hand off tasks from spotting a data signal to actually creating a visual design without us touching anything.

What we learned We learned that "Context is King." An AI can generate a t-shirt in seconds, but generating a good t-shirt requires giving the agent a very specific persona (like "Streetwear Hypebeast"). We also learned that prediction markets are a surprisingly accurate signal for consumer intent, not just gambling.

What's next for Prophet We want to use live data to adjust pricing as well to try and predict the best price for the products. For instants rising tensions with US and China could result in higher computer hardware prices. We want our system to be able to accurately predict them