Strike — Price Prediction Markets

Fully on-chain binary prediction market CLOB on BNB Chain with frequency batch auctions and Pyth oracle settlement.

Website: strike.pm · App: app.strike.pm · Docs: docs.strike.pm

---

What is Strike?

Strike is an on-chain orderbook for binary price prediction markets. Traders buy YES or NO outcome tokens on whether an asset's price will be above a strike price at expiry. Markets clear via frequency batch auctions (FBA) — uniform clearing price, pro-rata fills, no front-running. Settlement is trustless via Pyth oracle.

How It Works

1. Markets are created with a price feed, strike price, and expiry time
2. Traders place GTC (Good-Til-Cancel) or GTB (Good-Til-Batch) orders at a tick (1-99 cents)
3. Batches clear every 12 seconds — all orders in a batch get the same clearing price
4. At expiry, Pyth oracle determines the outcome — YES tokens pay $1, NO tokens pay $0 (or vice versa)
5. Winners redeem their outcome tokens for USDT

Key Properties

• 1 YES + 1 NO = 1 USDT — fully collateralized, no counterparty risk
• Batch auctions eliminate MEV and front-running
• Direct-from-wallet trading — no deposit/withdraw flow
• Permissionless redemption — winners claim anytime after resolution

Architecture

Smart Contracts (Solidity / Foundry)

Contract	Purpose
MarketFactory	Creates and manages market lifecycle (Open → Closed → Resolved)
OrderBook	GTC/GTB order placement, cancellation, and fill tracking
BatchAuction	Frequency batch auction engine — clears batches with uniform price
Vault	USDT escrow — locks collateral on order placement, releases on cancel/fill
OutcomeToken	ERC20 YES/NO tokens minted at fill, burned at redemption
Redemption	Post-resolution token redemption (1 winning token → 1 USDT)
FeeModel	Configurable fee structure (currently uniform 20bps)
PythResolver	Trustless market resolution using Pyth price updates

Infrastructure (strike-infra)

All backend services are written in Rust (axum, sqlx, alloy-rs, tokio).

Service	Purpose
Indexer	Indexes on-chain events into PostgreSQL
API	REST API for markets, orderbook, positions, referrals
Unified Keeper	Single binary running 4 tasks: batch clearing (3s), market creation (scheduled), resolution (5s), pruning (10s)
Market Maker	Automated liquidity provision with inventory-aware quoting

Frontend (strike-frontend)

Next.js + shadcn/ui + wagmi/viem. SIWE authentication, RainbowKit wallet connection.

SDK

The Strike SDK (strike-sdk on crates.io) lets you build trading bots and integrations in Rust. On-chain first — all live data comes directly from BSC via RPC/WSS.

use strike_sdk::prelude::*;

let client = StrikeClient::new(StrikeConfig::bsc_testnet())
    .with_private_key("0x...")
    .build()?;

client.orders().place(market_id, &[
    OrderParam::bid(50, 1000),
    OrderParam::ask(60, 1000),
]).await?;

📖 SDK Docs · 📦 crates.io · 📂 sdk/rust/

Development

Prerequisites

• Foundry (forge, cast, anvil)
• Rust 1.82+
• Node.js 22+
• PostgreSQL 15+

Build Contracts

cd contracts
forge build
forge test

Run Tests

# All 302 contract tests
forge test -vvv

# Specific test
forge test --match-test testClearBatch -vvv

Contract Addresses

BSC Mainnet (deployed block 95210316)

Contract	Address
MarketFactory	0x34E0BCC1619dBc6A00A23b70BbaD9F36b0483d82
OrderBook	0x71F7Bc523FFF296A049a45D08cBD39D39d3C047B
BatchAuction	0x9d66fa0Aad92bb4428947443c1135C06a0cbFFBb
Vault	0x43D5caC88a87560Db8040Bef16F0ce8871B4F7ee
OutcomeToken	0xdAA6810Ca9614e2246d2849Be2a9c818707e404B
Redemption	0xcC1687A27133f06dB96aF4e00E5bA91411f9c999
FeeModel	0xFd7538Ad9EFEe4fCE07924F65a30688044e0800C
PythResolver	0x3E0864BbC19ca92777BB4c2e02490fC0C7A44C5a
AIResolver	0x3e0D91480147802D9C41068d91b7878E7943a632
USDT	0x55d398326f99059fF775485246999027B3197955

BSC Testnet (deployed block 104337216)

Contract	Address
MarketFactory	0xB4a9D6Dc1cAE195e276638ef9Cc20e797Cb3f839
OrderBook	0xF890b891F83f29Ce72BdD2720C1114ba16D5316c
BatchAuction	0x743e60a7AE108614dDCb5bBb4468c4187002969B
Vault	0xb7dE5e17633bd3E9F4DfeFdF2149F5725f9092Fe
OutcomeToken	0x612AAD13FB8Cc41D32933966FE88dac3277f6d2a
Redemption	0x28de9b7536ecfeE55De0f34E0875037E08E14F88
FeeModel	0x78F6102Ee4C13c0836c4E0CCfc501B74F83C01CD
PythResolver	0x2a7fba2365CCbd648e5c82E4846AD7D53fa47108
AIResolver	0xE1C9DA3d9b00582951f25D35234F8580DE1646d9
MockUSDT	0xb242dc031998b06772C63596Bfce091c80D4c3fA

For the full canonical registry, including oracle and provider addresses, see docs/contracts/deployments.md.

History

The original prototype was built as a hackathon MVP using a parimutuel pool model with a Telegram bot interface. That code is preserved in the poc branch. The current version is a complete rewrite with a proper CLOB, batch auctions, and a web frontend.