HODL.fun
HODL the Future — One Prediction at a Time.
About
The problem it solves **The Problem it Solves — HODL.fun ** We’re bringing **pump.fun** vibes to the prediction market world — making it social, **tradable**, and fun, instead of boring, static, or gated. * Prediction markets are broken — they're boring, fragmented, and built for analysts, not internet natives. * Most prediction platforms are cold, spreadsheet-style apps with little user engagement. * They lack price discovery — one user = one vote systems ignore demand intensity. * There’s no social layer — no chats, no hype, no fun. * Newcomers find it hard to engage — no clear UX, and no incentive to stay. * No way to trade or cash out your conviction once you’ve made a choice. **Here’s what we’re solving:** ** 1. Prediction Markets Aren’t Designed for the Onchain Social Generation **Traditional prediction platforms feel like spreadsheets — they’re analytical, not entertaining. There’s no vibe, no community, and no easy way to get others excited. People want to express what they believe will happen — but in a fun, fast, and low-barrier way. ** 2. Tokenizing Beliefs: Every Outcome is a Market **We are tokenizing prediction markets — each outcome becomes its own token, complete with price chart, volume, and market cap. Outcomes are powered by bonding curves, so early believers can buy in cheap and exit profitably as hype grows. Users no longer just "vote" — they can buy/sell their belief, trade out, or double down. ** 3. Chat-Native Markets: Powered by XMTP + Base AgentKit **Every prediction runs inside an XMTP-powered chat, with a dedicated onchain agent running the room. These agents (built using **Base’s AgentKit**) act like moderators + market facilitators:  “2 hours left to join!” “Odds are shifting!” “Here’s your position summary.” It’s no longer a one-way action — it’s a conversation-driven market. ** 4. Culture-Driven Predictions (Not Just Finance or Politics) **Instead of dry economic topics, we let users bet on internet culture: “Will Solana flip ETH this year?” “Will Drake release an album in July?” “Will a new meme coin hit $100M in 48 hours?” It’s fun, hype-based, community-fueled, and shareable. ** 5. Built for Farcaster – The Onchain Social Layer **HODL.fun runs as a Farcaster mini app, meaning: Users can explore predictions like feeds See their profile with past positions Share predictions or charts to casts Social primitives are baked in — no need for a separate dApp. ** 6. Market Charts, Positions & Transparency **Every market has a live price chart showing how sentiment shifts. Users can view all positions they hold, how much they’ve gained/lost, and when they entered. It’s fully onchain, auditable, and real-time. ** 7. Feels Like Pump.fun — but for Predicting Outcomes **Inspired by pump.fun, we’re recreating that viral simplicity and flash, but for outcome markets. The UI is playful, the energy is high, and everything feels liquid and alive. ** 8. What’s Coming Soon **Live streaming of major markets for communal hype Leaderboards to track the top predictors Community-created prediction rooms with profit-sharing **Summary ****HODL.fun** is solving the cold, inaccessible nature of traditional prediction markets by turning them into living, tokenized, social experiences — where belief is tradable, agents run the show, and anyone can be early to the next viral moment.
How it's made
Challenges we ran into **Challenges I Ran Into ** **1. Integrating XMTP into a Dynamic Frontend Chat Flow ** One of the biggest challenges was setting up a smooth and responsive chat experience using XMTP, especially within a dynamic prediction market flow. XMTP is powerful, but: It requires careful identity management (wallet connection, message persistence). We had to ensure chats were scoped correctly for each game, user, and group — all while maintaining low latency. We also had to sync these interactions with our custom AgentKit-based agents that manage market logic and notifications. **We overcame this by: ** Building an abstraction layer to isolate XMTP message logic. Caching and indexing conversations locally for faster rendering. Using XMTP’s client.conversations.streamMessages() to trigger live updates inside each prediction room. **2. Connecting Smart Contracts to the Frontend (Especially Bonding Curves) ** Another core challenge was interfacing with the bonding curve smart contract: Bonding curve logic is highly state-dependent and gas-sensitive, so reflecting accurate pricing and slippage in real time was tricky. Syncing the frontend UI with contract state updates (e.g., token supply, pricing, pool reserves) introduced race conditions. **To solve this: ** We built an event-driven model using ethers.js to listen to contract events. We structured the contract functions to return clean, UI-friendly outputs (price previews, minimum received tokens, etc.). We optimized the read/write paths to reduce lag, and mocked contract responses during early development. **Deployed Smart Contract Address on Testnet** - 0x6BCcAb5A50b1D2116b18D40CCC5E845717961dee **Transaction Hash: 0x0ace84131e929b1aa5573f3ceb101bd951facbf118c84a5351f124ad2bb5098e **