Orba is a Telegram chatbot that lets you bet on predictions using natural language. No apps, no dashboards—just text. Say "I think Bitcoin hits $100k by June. Put $20 on it" and Orba handles everything: creates the market, executes your trade, tracks your portfolio, and pays you automatically when you win. Built on Celo with cUSD stablecoin settlement. Integrates ERC-8004 agent identity, Self AI human verification, and x402 payments.
How it works:
Technical architecture:
We're building the conversational interface layer for decentralized prediction markets. Our mission is to unlock the forecasting potential of everyday conversations by making it effortless to create markets, trade positions, and settle outcomes—all through natural language on the messaging platforms people already use.
Starting with WhatsApp and Celo, we aim to bring prediction market access to the 2+ billion users in mobile-first economies who are excluded from existing platforms by complexity and infrastructure requirements. If betting on your beliefs becomes as easy as sending a text, prediction markets can finally achieve the mainstream adoption needed to fulfill their promise as humanity's most accurate forecasting tool.
Prediction markets are one of the most powerful tools for aggregating information and forecasting outcomes. They consistently outperform polls, pundits, and models. Yet despite decades of development, they remain a niche product used primarily by crypto traders and finance professionals.
Three barriers block mainstream adoption:
Interface complexity: Existing platforms like Polymarket require users to understand order books, liquidity pools, and probability pricing. The average person doesn't want to learn a new financial interface just to bet $10 on an election.
Distribution challenges: Standalone prediction market apps compete for attention against Instagram, TikTok, and games. User acquisition costs are high, and retention is low because there's no existing habit to hook into.
Liquidity fragmentation: Markets need participants to function. New platforms struggle with the cold-start problem—no liquidity means bad prices, which means no users, which means no liquidity.
Meanwhile, the raw material for prediction markets already exists at massive scale. Every day, 2+ billion WhatsApp users debate predictions in group chats: sports outcomes, crypto prices, election results, company earnings, weather, celebrity gossip. These conversations represent billions of dollars in latent prediction market activity—opinions that people would bet on if betting were as easy as typing a message.
Orba eliminates every friction point between having an opinion and putting money behind it.
Conversational interface: Instead of learning a trading platform, users just talk. Orba's GPT-4 agent understands natural language like "I bet Arsenal wins tonight" or "What's the price on the Bitcoin market?" and translates it into on-chain actions. The learning curve is zero because everyone already knows how to text.
Telegram distribution: Rather than building yet another app, Orba lives inside the messaging platform people already use daily. Adding Orba to a group chat instantly turns that group into a prediction exchange. Friends who argue about sports or crypto can now settle debates with real stakes. This creates viral growth—every group that adds Orba becomes a distribution channel.
Built-in liquidity: The LMSR market maker provides instant liquidity for any market size. Unlike order book systems that need matching buyers and sellers, LMSR lets a single user trade immediately at mathematically fair prices. Social groups provide natural liquidity pools—the same people who debate predictions are the ones who'll bet on them.
Automated resolution: Oracle integrations remove the need for manual settlement. Crypto price markets resolve automatically via CoinGecko data. Sports markets resolve via API-Football. Users bet, wait for the outcome, and receive winnings without any action required.
Custodial simplicity: Each phone number gets a deterministic HD wallet. Users don't manage keys or sign transactions—they just text. For the hackathon demo, this runs on Celo Sepolia testnet with test cUSD. Production would add optional key export for users who want self-custody.
The result: prediction markets that feel like group chat games, not financial instruments.
Transaction fees on prediction market trades (0.5-2% per trade), plus premium features for power users and group administrators. Revenue scales directly with trading volume.
Want to grow through just grant funding
Global, with primary focus on mobile-first emerging markets including Sub-Saharan Africa, Latin America, Southeast Asia, and South Asia—regions where WhatsApp dominates messaging and mobile money has already normalized phone-based financial transactions.
0 USD
