소개
Decompose provides AI agents with a crucial preprocessing step, dissecting raw text into meaningful, classified semantic units like authority, risk, attention, and entities. This tool operates deterministically without relying on Large Language Models (LLMs), ensuring speed, consistency, and offline capability. By structuring textual information before it reaches an LLM, Decompose significantly reduces token consumption, enhances relevance in retrieval-augmented generation (RAG) systems, and enables intelligent routing of content based on its importance and type, making agent workflows more efficient and precise.
