Origin & Motivation

Why EKOS was built.

EKOS started from a repeated experience on legacy estates: the most valuable knowledge — what a pipeline really does, why a table exists, which service will break if a column changes — was never written down. It lived in code, stored procedures, ETL jobs and people's heads, and each new engineer (and now each new AI agent) had to rediscover it by reading.

Two ideas shaped the project:

• The enterprise already contains its own documentation. The gap is not information but a compiler that can extract it reliably, the way a language compiler extracts meaning from text.
• Trust needs evidence. Fluent answers are cheap; verifiable ones are not. So every conclusion in EKOS must be traceable to where it came from, and the store must be append-only so nothing is silently rewritten.

The project is developed RFC-first (every capability is designed in writing before it is built — see RFC Process) and documented in a running series of devlogs in the repository. It is written in Rust and licensed under MIT.