The command centre for a closed-loop irrigation fleet.
The Intelligence Platform is the dashboard this whole site points to: live telemetry from real field hardware, irrigation and fertigation control with server-verified commands, alerts, and a savings ledger. The platform runs today against our bench pod, built to scale to a fleet.
One dashboard, from probe reading to pump command.
Every module below runs today, against the same bench pod whose telemetry powers the ribbon on our home page. Nothing here is a mock-up.
Live telemetry Live
Soil moisture, temperature, EC, pH, N/P/K, air quality and camera frames, refreshed on every reporting cycle from the field pod.
Irrigation control Live
Start, stop or override a watering cycle from the dashboard, with the confirmation and acknowledgement flow described below: no silent commands.
Fertigation & pH Building · dosing logic
Nutrient dosing and pH correction commands, issued through the same control channel and logged against the same acknowledgement semantics as irrigation.
Alerts Live · rules
Threshold and connectivity alerts surface in the dashboard as soon as a reporting cycle is missed or a reading crosses a configured bound.
Savings ledger Live · metered maths
Litres and cost saved against a metered baseline, with the formula and gaps disclosed: see how we measure impact.
Pilots view Building
A fleet-level view across pilot sites as they come online: see how the pilot programme works.
A dashboard button never talks to a valve directly.
Every irrigation or dosing command is signed and issued from the server, so the signing key never reaches the browser. Each command carries a monotonic sequence number, and the hub will not act on a command it has already seen or one that arrives out of order.
The hub acknowledges every command it receives (execution or rejection) before the dashboard marks it complete, so an operator always knows whether a command actually reached the field. A one-shot manual watering command is capped at five minutes of pump run-time regardless of what was requested, a hard limit enforced independently of the confirmation step.
| Server-side signing key (never in the browser) | Enforced |
| Sequence number + acknowledgement per command | Enforced |
| 5-minute cap on one-shot manual watering | Enforced |
| Out-of-order and duplicate command rejection | Enforced |
| Role-based multi-operator permissions | Planned |
{
"cmd_id": "c-88213",
"seq": 4821,
"action": "irrigate",
"zone": "Z3",
"signed": true,
"duration_s_cap": 300
}
→ hub ack: { "received": true, "executed": true,
"duration_s": 240, "seq_ok": true }
See it running on real hardware.
The dashboard is live against our bench pod in Kuwait right now. Open it, or read how the hardware and edge AI underneath it work.