llm-context-search is a fast, LLM-free engine that turns web search results into clean, ranked, token-bounded context for AI agents.