I turned a YouTube trading video into a backtested strategy — with Claude + MCP
A developer used Claude and a tool called Scribefy to extract a trading strategy from a YouTube video transcript. The process involved using Claude to analyze the transcript and identify the core mechanics of a "Power-of-Three (PO3) dealing range" strategy. This logic was then translated into a backtesting script, which revealed the original idea was marginal but led to a refined session-based breakout variant that showed promising results on gold. AI
IMPACT Demonstrates how LLMs can be integrated into specialized workflows to extract and analyze information from unstructured data like video transcripts.