VATES is an embeddable B2B AI knowledge platform designed to help organizations manage specialized, frequently changing, and non-public information.
Unlike conventional vector-based RAG systems, VATES uses EchoScript (ES), a structured intermediary representation, combined with BM25 and heading-path retrieval to deliver traceable answers without a vector database.
Users can create knowledge structures through natural language, refine them manually, and deploy AI assistants through widgets, standalone interfaces, or APIs.
Built with Python, FastAPI, React, and AWS, VATES focuses on knowledge ownership, accuracy, and traceability.
