Show HN: Strot – Compile an LLM's intent once, run decisions locally forever

Strot is an open-source tool that compiles an LLM's decision logic once during a setup phase, then runs lightweight local decision-making on production data without requiring further LLM calls. After initial compilation with an LLM, it generates a frozen Python artifact that classifies, scores, or routes records locally at minimal cost with bounded outputs and confidence thresholds that allow abstaining on uncertain decisions.