Compress and minimize LLM context. Turn verbose, messy, unstructured text into a dense token-minimal JSON object that keeps every fact. Reduce token usage, shrink prompts, and cut context costs for AI agents.