Descriptions:
This video explores using Claude Code as an autonomous trading-strategy portfolio manager, connecting it to a custom REST API that exposes a library of 2,870 backtested and forward-tested trading strategies built over five years on the Strategy Factory platform. The creator gives Claude the ability to read all strategy statistics, activate and deactivate live bots on Bybit (which provides a native MCP server), and make portfolio allocation decisions without manual input.
The setup involves connecting a Bybit exchange account via API keys, granting Claude access to the Strategy Factory API endpoints — including strategy lists, performance data, and bot activation toggles — and providing a detailed prompt instructing the model to act as a quantitative hedge fund. Claude Code then writes a full Python application from scratch, including an HTML dashboard that updates on a configurable schedule (hourly, every 8 hours, or daily) and implements Markov regime detection to identify whether current market conditions favor trend-following strategies or suggest staying in cash.
The creator shares the full prompt used, describing it as a playbook viewers can adapt to their own platforms. The broader concept — delegating not just strategy coding but live strategy selection and bot lifecycle management to an AI agent — represents an emerging pattern of using coding agents as operational infrastructure rather than purely as development assistants. The video concludes with a live recording of Claude Code building the entire system in real time.
📺 Source: Trading with DaviddTech · Published April 06, 2026
🏷️ Format: Hands On Build
