Show HN: Clawock – AI argues, code settles the trades

Category: ai-ml

Tags: trading, ai-agent, decision-ledger, finance, plugin-architecture

Score: 7.0/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 6)

Clawock is a framework that lets AI agents argue bull and bear cases for stock trades, while code independently settles the outcomes against market prices to prevent self-grading bias. It provides a plugin architecture for indicators, a ledger of settled decisions, and a recursive self-improvement loop where past outcomes inform future judgments. The project is interesting for its rigorous separation of AI reasoning from deterministic settlement and for its live, real-money track record (with losses included).

Target audience: quantitative traders, AI engineers, fintech developers

Repository: https://github.com/KCNyu/clawock · Python · MIT · 21 stars

View on Hacker News