Rules written into the code
These are enforced by the software, not left to a policy document.
Local-first AI
Inference runs on hardware at the hub. A cloud AI model would only ever be added by an explicit, disclosed decision.
Simulated means labelled
Drones, sensors and partners that don't exist yet sit behind real interfaces and carry a "simulated" tag on screen.
Verify before treating
Imagery can flag a zone, but nothing is sprayed until someone confirms the problem on the ground.
The farmer decides
The AI can suggest clearing a thicket or felling a tree. The farmer makes the call, and shea and other economic trees are never removed by default.
More rules the platform enforces
The same idea, applied to the field, the money and who may change what.
Advice by zone
Guidance is always for a zone of the field, never the whole field, and treatment covers only the zones that need it.
The right index
Germination flights use OSAVI only, because the canopy is too small for NDRE. The software enforces it, not just the screen.
A proxy stays a proxy
Soil conductivity is shown as an estimate, never labelled as nitrogen, phosphorus and potassium.
No waiting on the internet
Inference never blocks on a connection. The models run where the data is.
One ledger
Every amount moves through an immutable ledger. Balances are computed from it, never edited.
The hub approves
Drones, equipment and lab tests go out only after the hub approves them, and are charged at approval.
Fully funded first
A loan can't be paid out until it is fully funded, and harvest sales repay it automatically.
Read-only means read-only
Investor and government logins can look at everything and change nothing: the server refuses any change they try to make.
Put the engine to work with us
For partners, investors, government programmes and operators that want to run a pilot. Your message goes straight to the Aionautics team.










