Last released Sep 21, 2026
Deterministic, local-first context optimization for LLM prompts: clean, dedupe, structure, and measure token and cost savings before calling a large model.