02 — How it works
Four steps from facility to schedule.
A read-only integration. A continuous data collector. A constraint solver. A schedule your operations team can trust. No changes to workload code.
01
Step 01 · Connect
Configure your facility.
Configure your facility — solar capacity, battery size, IT load, diesel cost.
A guided setup captures your PV nameplate capacity, battery chemistry and depth-of-discharge limits, average and peak IT load, and the marginal diesel cost per kWh. Read-only credentials are issued for your BMS and workload queue.
02
Step 02 · Collect
Pull live signals.
Gridient pulls live solar forecasts, carbon signals, and battery state every 5–15 minutes.
Numerical weather prediction feeds your generation forecast. Carbon intensity is fetched from regional providers, or computed from our diesel proxy where coverage is absent. Battery telemetry comes from your BMS over MQTT or Modbus.
03
Step 03 · Optimise
Solve for the next 24 hours.
The CP-SAT solver finds the provably optimal schedule for the next 24 hours.
We model jobs, dependencies, deadlines, and SLA penalties as a constraint program. Google's OR-Tools CP-SAT solver returns a provably optimal schedule in median 23 milliseconds, even with 200+ jobs and complex battery state-of-charge constraints.
04
Step 04 · Save
Run jobs in solar windows.
Jobs run in solar windows. Diesel spend drops. Emissions fall. SLAs hold.
Beta deployments are reporting a median 47% reduction in diesel consumption and 41% lower gCO₂/kWh, with no regressions on SLA timing. The schedule is delivered as a webhook your orchestrator already trusts.
From read-only credentials to a working schedule in under a week.
Most beta facilities go from initial connection to a fully optimised 24-hour schedule within five working days.
Start the conversation →