About weathership
Open-source AI tooling for enterprise data systems.
weathership builds open-source AI tooling for the systems that run enterprise — data warehouses, knowledge graphs, manufacturing systems, and the federated compute fabric that ties them together. Each project takes a system in motion, reads the conditions that define it, and turns that reading into structured signal — classifications, belief intervals, traces, alerts — that a downstream operator can act on.
What we build
The portfolio spans four domains: metadata governance for enterprise data warehouses (Signals, Aegir, Atelier, SDG corpora), graph-oriented exploration (Gaius), event-sourced manufacturing systems (Reach), and a global resource-oriented compute fabric (Vigil). Each project is observable in production, inspectable in source, and documented as it ships. We don't make claims we can't back up with code and measurement.
Mission
To build durable, transparent AI tooling — systems whose behavior is inspectable, whose decisions can be traced, and whose failure modes are documented before they're hit. We release as open source by default because the systems we want to exist need to be reviewable by the people who depend on them.
Values
- Open
- The software ships under permissive licenses. The architectural decisions are written down. The trade-offs we made — and the ones we didn't — are visible to anyone who wants to read them.
- Transparent
- When we say a system does something, it does that thing. When it doesn't, the gap is named. We don't paper over uncertainty with confidence we haven't earned.
- Durable
- We optimize for the system that's still useful in five years over the demo that's compelling next week.
Origin
A weather ship was once a vessel stationed in open ocean for the express purpose of gathering atmospheric data — a floating instrument that turned uncertainty into signal so the ships following behind could plan their crossing. Our work plays the same role for the systems we build alongside.
weathership was started by Ryan Hill in 2026 to hold several related strands of work in open-source AI for enterprise data systems under one umbrella. The projects predate the umbrella; the umbrella exists because they benefit from being held together.
Press, brand, and assets
The full media kit has the logo (SVG + PNG renders), color palette with copyable tokens, type system, usage guidelines, and press boilerplate. Everything on this site is downloadable.