Description

VATES is an embeddable B2B AI knowledge platform developed by STANDOUT in Japan. It enables organizations to build AI assistants that provide traceable answers based on company-controlled information.

Unlike conventional vector-based RAG systems, VATES uses a structured intermediary representation called EchoScript (ES), combined with BM25 and heading-path retrieval. This approach enables knowledge to be organized, reviewed, and updated without relying on a vector database.

Users can create structured knowledge through natural language, refine records manually, and deploy AI assistants through an embeddable widget, standalone interface, or API.

Key capabilities include role-based access control, audit logs, configurable data retention, CSV import, and multilingual support.

VATES is designed for organizations managing specialized, frequently changing, or non-public knowledge where accuracy, ownership, and traceability are essential.