dr-manhattan is an open source unified API designed specifically for prediction markets, functioning similarly to how CCXT works for cryptocurrency exchanges. It provides developers with a standardized interface to interact with multiple prediction market platforms simultaneously, solving the liquidity fragmentation problem across different markets.

The product offers a unified interface that allows writing exchange-agnostic code working across all prediction markets. It includes WebSocket support for real-time market data streaming with built-in connections for live orderbook and trade updates. The platform features a strategy framework with base classes for building trading strategies, order tracking, position management, and event logging. It provides full type safety with comprehensive type hints throughout the codebase and standardized order management capabilities for creating, canceling, and tracking orders with consistent error handling across exchanges.

The system works by providing abstract classes and methods that developers can implement to add new exchanges, making integration straightforward through clean architecture. The API follows a consistent pattern where developers initialize exchanges with the same interface and can fetch markets and perform operations using standardized methods regardless of the underlying platform.

Benefits include simplified development workflows, reduced code duplication, and the ability to deploy applications across multiple prediction markets simultaneously. The tool enables developers to build sophisticated trading applications, market-making bots, and analytical tools that can operate across the fragmented prediction market ecosystem.

dr-manhattan targets developers building applications for prediction markets, particularly those working with platforms like Polymarket, Kalshi, Opinion, Limitless, and Predict.fun. It's designed as an AI-native solution built specifically to address the liquidity problem in prediction markets through standardized API access.