About ATAR
Precision spraying.
Built in the field.
We’re a team of engineers, agronomists, and farmers building the next generation of smart spraying technology — from Novi Sad, Serbia, for fields worldwide.
Our mission
Reduce herbicide use by 80%. On every pass.
Agriculture accounts for billions of liters of herbicide sprayed annually — most of it landing on bare soil. ATAR changes that by activating nozzles only when and where a weed is detected.
Our systems use onboard AI to process 60 frames per second, classify plants in under 42 milliseconds, and control every single nozzle independently. No cloud dependency. No dead zones. Just precision.
Novi Sad · Serbia
By the numbers
Built for real fields, not trade shows.
60 FPS
Camera capture rate
<42ms
Detection latency
80%
Herbicide reduction
40+
Compatible brands
What drives us
Engineering principles
Field first
Every decision is tested in real conditions — dust, heat, vibration, 14-hour days. If it doesn’t survive a season, it doesn’t ship.
No cloud dependency
Every module has its own GPU. The system works offline, in fields with zero coverage. Your data stays on your machine.
Open integration
CAN bus, ISO-BUS, standard nozzle fittings. We work with your existing equipment, not against it.
Honest specs
We publish real numbers from real fields. No lab conditions, no asterisks, no “up to” claims without the data behind them.
Farmer-configurable
Lead time, trail time, nozzle count, spray pressure — every parameter adjustable from the cabin. Your field, your rules.
Continuous learning
Every field we operate in makes the AI smarter. Regional weed models improve with every season of data.
Our journey
From prototype to production
First prototype — Single-camera module tested on a 12m boom in Vojvodina. Proof of concept for real-time weed detection at field speed.
Multi-module architecture — Scaled to full boom coverage with independent GPU per module. Pneumatic nozzle control integrated.
Pressure regulation — ARAG proportional valve integration for constant dose delivery. First commercial pilots with partner farms.
Green-on-green detection — AI models trained to distinguish weeds from crops, not just green from brown. 3D configurator launched.
Production scale — Full product lineup available. Customer portal, OTA updates, and Intelligence SaaS subscription live.