Kanopi wins AgriSpark 2.0 - AI Edition hackathon organised by the German-Thai Agri Cooperative GETHAC and the Dept. of Agricultural Extension Thailand .
The signals that decide a season are present in the field long before they are obvious. Kanopi runs crop-specific models on the hardware farms already own (the phones in workers' pockets, fixed field cameras, drones), doing early pest and disease detection, fully on-device. Geotagged field observations flow into one centralised operational dashboard, so plantations, cooperatives, corporates and national programmes can act with complete information.
Why it matters
Up to 40% of crop production is lost to pests and disease each year, much of it preventable. The earliest signs are often specific and legible: a discoloured leaf, a sesame-seed-sized lesion, a thinning strand, or a stray insect. The challenge is not recognising these signals. It is observing them consistently enough, across enough ground, to act before damage is done.
Most agricultural systems were not built for this. Manual scouting samples only a small portion of the canopy, satellite imagery often lacks the resolution needed to identify emerging threats, and operational software is largely concerned with administration rather than understanding field conditions in real time. As a result, critical information remains trapped at ground level until it surfaces as yield loss, disease spread, unnecessary input expenditure, or supply disruption.
The consequences differ by scale. On a plantation, it may be a soil-borne disease discovered after containment is no longer practical and losses have already been locked in. For a cooperative, it may be an outbreak that spreads faster than agronomists can respond. For a corporate buyer, it may be a supply shortfall that becomes visible only after the opportunity for intervention has passed.
How it works
Agricultural decisions are only as good as the information behind them. The challenge is not simply detecting a problem, but collecting observations at scale, turning them into an operational picture, and extending coverage into environments where conventional approaches break down.
The detection layer.
Spore is a platform-agnostic runtime that runs crop-specific, client-tuned models directly on commodity hardware, without a cloud round trip. Every detection is geotagged and pushed upstream into Toya, and optionally into existing client infrastructure. Our mobile app is just one Spore modality (50 to 100 millisecond inference on a $90 Android A07), the same core compiles onto fixed field devices and drones.
The result is a continuous stream of field-level observations generated where problems first emerge, rather than where they eventually become visible from above.
The aggregation layer.
Individual detections are useful, but patterns across thousands of detections are the true operational value. Toya integrates heterogeneous field signals, including RGB, multispectral, LiDAR and acoustic inputs, alongside weather and satellite data, into a single operational view.
Managers gain a live picture of disease pressure, crop health and emerging risks across an entire operation, while the underlying intelligence improves with every season and every new field observation.
The hardware frontier.
Extended-range, repairable autonomous drones and robots with rugged edge compute for harder applications such as long transects between plots, aquaculture and difficult terrain.
Hardware, software and AI are typically purchased as separate systems and stitched together through integration work. Kanopi develops edge detection, operational software and custom hardware within a single architecture, allowing organisations to deploy individual components where specific gaps exist or adopt a fully integrated solution from detection through to decision-making.
Built for
An estate's commercial life depends on keeping as many productive blocks in rotation as possible, yet its most serious threats often develop out of sight. Ganoderma reaches the majority of mature oil palm estates across Southeast Asia, spreads underground through root contact, and has no cure once established. By the time symptoms are visible at scale, containment is often no longer practical and losses have already been locked in.
Kanopi gives estate managers a systematic way to find problems before they become operational realities. Spore runs on the devices field crews already carry and can be tuned to detect early signs of disease, stress or deficiency across every block. Those observations flow into Toya, creating a live spatial picture of pressure across the estate. Decisions on treatment, containment and replanting are then guided by what is actually happening in the field, allowing operators to protect productive acreage, deploy resources more efficiently and extend the economic life of the estate.
A cooperative creates value when it helps members achieve outcomes they could not achieve alone. In crop protection, that means turning hundreds or thousands of fragmented farms into a shared intelligence network. A planthopper outbreak can collapse a rice territory in under three weeks, a blast outbreak in under two, yet most cooperatives have limited agronomic capacity spread across vast areas. Manual scouting does not scale, while blanket recommendations waste inputs by treating every field as equally at risk.
Kanopi allows cooperatives to direct attention where it matters most. Through Spore, every member contributes field-level observations with no specialist training required, while Toya aggregates those detections into a live operational picture of risk across the cooperative's territory. Agronomists can see pressure building, prioritise intervention where it will have the greatest impact, and deliver targeted guidance before losses spread. Members spend less, lose less, and receive advice grounded in conditions on their own farms, giving the cooperative a tangible service that individual growers cannot replicate alone.
For organisations sourcing through contract farmer networks, the central challenge is visibility. Disease outbreaks, quality issues and production shortfalls rarely appear as field problems. They appear later as procurement problems, often after there is little left to influence. Regulatory requirements such as EUDR add another layer of complexity, demanding farm-level traceability across supply chains that were never designed to produce it.
Kanopi turns fragmented sourcing regions into an operationally visible supply base. Spore deploys directly onto farmer devices with no dedicated hardware rollout, generating geotagged field observations across the network. Toya aggregates those signals into a single operational view, giving procurement teams earlier warning of emerging risks and more time to respond. The same plot-attributed, timestamped record also creates traceability as a by-product of day-to-day operations, allowing compliance requirements to be met through normal field activity rather than separate reporting exercises.
Pricing
Kanopi enterprise is licensed annually to the operator org, priced by coverage and crop type; growers and outgrowers are included under the organisation's coverage.
Independent smallholders access Spore via the app at no cost on plots under 3 ha.
Partner with us
Request a demonstration, discuss API access, deployment support, or how Kanopi fits your operation.