Markets{dot}fun

Problem Statement

Prediction markets are booming. What started as a niche corner of crypto is now becoming a mainstream way to bet on the future, from politics to sports to crypto events. Every day, millions of users make predictions, track outcomes, and try to spot the next big trend. Yet, for all their popularity, it is all gambling if you know nothing about the market you are betting on. You might know nothing about Politics, but someone has an insane winning rate in that market. Currently, there is no way to copy that person's tradeThat’s where our platform comes in. We’ve built a social layer on top of prediction markets, turning isolated trades into a thriving network of insight. We built leaderboard with a real-time spotlight on top performers. Users aren’t just numbers on a chart; they’re ranked by their wins, profits, and success across different market categories. You can see who’s dominating today, who’s consistent over the week, and who’s building a track record over the last month.But it doesn’t stop there. Every user has a personal profile, showing their journey, analytics, and trading style. Soon, users can follow each other, turning the platform into a social space where strategies are shared, insights are discussed, and success becomes visible. The goal is simple: learn from the best, and maybe even become the best yourself. How we see the Prediction Market panning out is big creators, influencers will be coming onboard with their subject expertise on certain markets - their followers will come along and benefit from the trade that influencer hasWe’ve also added a copy trading layer. Imagine spotting a top trader, clicking “follow,” and automatically mirroring their trades. It’s mentorship in action—data-driven, transparent, and trustless. Beginners can learn faster, and top performers are rewarded with influence and visibility.Under the hood, the platform runs on real-time prediction market data from Polymarket, smart contracts for secure and verifiable actions, and analytics that turn raw data into meaningful insights. And for the curious minds, our AI layer quietly identifies market trends, helping users discover new opportunities and track emerging talent.In short, our platform transforms prediction markets into a social, interactive, and educational experience. It’s where data meets community, where traders learn from each other, and where prediction markets finally feel alive—not just as bets, but as a network of people shaping the future.

Solution

How We Built It – The Nitty-GrittyWe started with the vision of turning prediction markets into a social network of traders. To make that real in a hackathon timeframe, we stitched together several moving parts, from data pipelines to smart contracts to the frontend. Here’s the breakdown:Market Data & Analytics LayerPolymarket Substreams & APIs were our data backbone. Instead of reinventing the wheel, we tapped into Polymarket’s substream packages to fetch PnL and user trading history.We processed this data into daily, weekly, and monthly profit percentages as well as category-wise win rates.This was the fuel for our leaderboard and profile pages.(Hacky thing we did:) Since we didn’t have time to build a full custom indexing solution, we wrote quick scripts to filter and aggregate the substream data into just the analytics we needed (1d/7d/30d).Smart Contracts & Copy TradingWe wrote our own copy trading contract that allows users to link their wallet to a trader they want to follow.Whenever the trader places a bet, followers’ wallets can automatically mirror it.We introduced a profit-sharing model: a small percentage of the follower’s profit flows back to the trader they copied, and the platform can optionally take a tiny cut.(Hacky thing we did:) To save time, we piggybacked on standard ERC20 transfer hooks and mirrored trade payloads instead of building an entirely new execution flow.Wallet & Payments IntegrationFor Web3 login and transactions, we used RainbowKit + Wagmi for smooth wallet connection.We also integrated Polygon for low-fee transactions, since many micro-copy trades and fee distributions would be prohibitively expensive on mainnet.For the bounty requirement, we tested PayPal’s PYUSD as an optional settlement token, making our UX more mainstream-friendly.AI Agent (x402 Chatbot)We built an AI-powered agent that analyzes live market conditions and suggests the best trades.Every time a trader places a bet through the chatbot, we charge a tiny fee, making the AI both useful and monetizable.We trained the agent on past market categories and structured it to recognize trends (politics, sports, crypto, etc.).(Hacky thing we did:) Since we couldn’t build a huge ML pipeline in time, we combined a lightweight rules engine with an LLM-based wrapper — just enough to look “smart” and deliver helpful suggestions quickly.Frontend Built with React + Next.js, our frontend shows: The leaderboard with sortable stats. Profiles for each trader (performance analytics + followers). Copy trading toggle for followers. We focused on visual clarity — the UI makes complex stats (PnL, ROI, win rates) look simple and intuitive for users.Backend Built using GraphQL's substream, Polygon Amoy x402 and Node.js