OpenAlice is a local trading workspace for AI agents that automates research and trading across equities, crypto, commodities, forex, and macro markets.
The tool addresses the mismatch between how coding agents work and how trading typically operates. Coding agents thrive in structured environments with git repositories, issue tracking, file systems, and review workflows. Trading, by contrast, usually involves scattered research, broker accounts, and ephemeral chat histories. OpenAlice maps trading work onto the collaboration substrate that coding agents already understand: workspaces with directories and git repos, markdown-based issues for self-describing tasks, a memory graph of tracked entities like assets and theses, an inbox for durable reports, market data tools exposed through CLIs, and a git-like approval workflow for account actions. This allows agents to research, analyze, propose positions, and execute trades within a shared, inspectable system.
Adoption suits traders and investors who want to delegate research and position management to AI agents while maintaining control through approval gates. The tool works for read-only research without a broker account, making it accessible for learning and backtesting before live trading. It complements rather than replaces existing coding agents; it provides them a trading-shaped workspace. The README emphasizes that the trading layer is experimental and beta, and explicitly cautions against live trading with real funds without full understanding of the risks.
Development shows active iteration on core features. The project maintains public communication channels and documentation. The codebase is written in TypeScript and remains under active development with incomplete interfaces and subject to breaking changes.
