Entroly drops Rust-powered LLM context optimizer

Entroly represents a shift from "brute-force" RAG to information-theoretic context management, proving that prompt efficiency is as much a systems problem as it is a linguistics one.

• –Rust core adds <10ms overhead per request, outperforming Python alternatives by 100x
• –Uses Shannon entropy scoring to rank "surprisal" in code, ensuring high-value logic is prioritized
• –Native MCP support enables seamless integration with Cursor and Claude Code without changing workflows
• –Knapsack-optimal budgeting dynamically fits the most valuable snippets into fixed token limits
• –Transparent proxy mode allows optimization for any AI tool by intercepting standard HTTP requests