Weather.Cal
Builds probabilistic weather intelligence with regional context, evaluation against observations and explicit uncertainty for downstream agents.
Weather.Cal profileAgent network
Focused intelligence. Shared discipline.
No single model should pretend to understand every domain. Our agents are narrow by design and connected through explicit, reviewable data products.
Power.Cal
Will connect weather intelligence with renewable generation, demand, grid context and system balance to study price-relevant energy scenarios.
Power.Cal profileMacro.Cal
Reserved for future macroeconomic research. Its scope, evidence contract and evaluation gates will be defined before implementation.
Domain reservedA repeatable contract
New agent, same standard.
Every future agent must define what it predicts, which evidence it accepts, how uncertainty is represented and which reviews can stop a decision. Shared infrastructure does not mean shared shortcuts.
- 01
Domain identity
A bounded subject, explicit outputs and independent ownership.
- 02
Evidence contract
Versioned inputs and outputs with a reproducible cutoff.
- 03
Governance policy
Required checks, veto rights and reasoned abstention.
- 04
Validation path
Research first, paper evidence next, any later mode separately approved.
The first producer-consumer path
Weather.Cal is the foundation. Power.Cal is the first downstream test.
The goal is not to duplicate weather logic inside an energy agent. Power.Cal should consume a governed Weather.Cal product, then own the additional uncertainty of generation, demand and market response.