Vibe-Trading is an open-source personal trading agent with 36,000+ stars on GitHub that gives an LLM agent a comprehensive set of trading capabilities in one command.
Developed by HKUDS, Vibe-Trading turns a general AI agent into a quantitative research and trading assistant. One command wires the agent into market data, strategy backtesting and trading workflows, so you can describe what you want in natural language and let the agent run the analysis. Under the hood it is a Python system with a FastAPI backend and a React frontend, distributed on PyPI as vibe-trading-ai. Multi-agent setups and MCP tool integrations let it pull data, test hypotheses against history and report results in a structured way. The repository is very actively maintained, with updates in October 2026. Like all trading automation, it is a research and tooling project — not financial advice.
The project comes out of a research lab rather than a brokerage, which shapes its design: strategies are things you can inspect, reproduce and share, and the agent’s reasoning is visible instead of hidden inside a black-box signal service. It is equally at home answering questions — why did this strategy draw down last quarter? — as it is executing a defined workflow on a schedule. Because everything runs under your own accounts and API keys, no third party ever holds your brokerage credentials or sees your positions. That local, open approach has made it a popular foundation for the growing wave of agent-driven finance experiments.
Vibe-Trading is free and open source under the MIT license. Your only costs are any LLM API usage and market-data or brokerage services you choose to connect.
Quant-curious developers, independent traders and researchers who want an agent that can fetch data, backtest ideas and automate analysis under their own control.