Sense the meter. Reason about load. Act within your limits.
A Fibric operator reads the meter and the grid, reasons about when to shift load and how to answer a demand-response event, and acts inside the limits you set. Every move is single-flight, reversible, and fail-closed. Governed, on your real data.
The meter, the grid, your resources, and the weather.
Connectors pull meters and sub-meters, grid and price signals, your distributed energy resources, and the weather into one canonical event. A battery's state of charge, a five-minute price, a load-shed window, an incoming heat wave: each lands as the same kind of signal the operator can reason over, on your real data, never a placeholder standing in for a reading.
When to shift, where a fault is forming, how to answer.
The same governed loop every Fibric operator runs, pointed at energy.
Move load off the peak
It reads price and demand together and proposes when to pre-cool, charge the battery, or back off the peak, with the reasoning shown in plain language on your real load.
Catch the drift early
It watches a sub-meter pulling more than its history, a circuit drifting out of band, a resource that stopped reporting, and flags where a fault is forming before it trips.
Meet the DR event
When a demand-response event arrives, it proposes the curtailment that meets the target inside your guardrails. The base model proposes the plan; a deterministic executor disposes.
Curtail, shift, or dispatch. Fail-closed at every limit.
It curtails a load, shifts it in time, or dispatches a battery, single-flight per entity so a signal can never fire the same action twice. If a limit is unclear, it holds instead of guessing.
Capability points to a connector through indirection, so swapping one meter vendor or DERMS for another is config, not a rewrite. Every dispatch writes an attributable, reversible receipt: who proposed it, which limits it honored, and exactly how to undo it.
A site that answers a DR event on its own.
Inside guardrails it can prove. Connect one meter and see your own load in an afternoon. Pay for the integrations you run.