Zero-maintenance pellet detection, appetite prediction and reliable infrastructure – designed for land-based RAS and sea-based production.
End-to-end feed optimization – from pellet detection to direct feed control integration.
Real-time pellet detection in return water and in-tank environments. Models tuned for turbidity, bubbles and dynamic lighting – delivering accurate counts regardless of conditions. The signal feeds directly into appetite models and control systems.
Short-horizon appetite models built on pellet signal, feeding history and tank context. Operators and control systems get actionable recommendations – when to increase, maintain or reduce feeding – updated continuously throughout the production cycle.
Edge-first AI deployed on-site with remote monitoring, automated model rollouts and standard APIs. Integrates with existing feeding and control systems via OPC-UA, MQTT and REST – no cloud dependency required.
Direct signal output to SCADA and feeding systems – not just dashboards. Detection and appetite data close the loop, enabling automated or semi-automated feed adjustments in real time.
Built around operational reality: noisy water, demanding environments and limited time.
Make advanced aquaculture AI feel simple, reliable and transparent – so operators and systems can trust and act on the signals. Every farm should have access to real-time feed intelligence, regardless of scale or technical resources.
Deliver pellet and appetite insight that directly improves feeding strategies, fish welfare and production stability across sites. We build solutions that work in the field – not just in the lab – and integrate with the systems operators already use.
We work directly with farmers and integrators to understand real operational challenges. Every feature is validated in production environments before release, and our subscription model ensures continuous improvement without additional cost.
A non-contact optical system for return water that uses advanced optics to detect and track individual pellets and waste particles – without touching the water. A tested, working solution with direct feed control integration. Install it once and it just works, with no cleaning or maintenance required.
Our system operates without touching the water, eliminating cleaning requirements and ensuring long service life. Install it once – it just works.
OPC-UA and MQTT providing real-time and calculated values to any modern automation system.
With Dynamic Open API, aggregated data and pre-calculated values are available to third-party systems that cannot use OPC-UA or MQTT directly.
Models are continuously retrained and deployed. Your system improves over time with no effort from your side.
We believe big savings should come cheap!
Example processing of a return-water stream with pellet detection and tracking.
Calculation for 20 T feed per day.
| Scenario | Spillage | ΔC spill | ΔC FCR | ΔC saving |
|---|---|---|---|---|
| Reduced Spillage (1%) | 1.00 % | 1 440 000 kr | 0 kr | 1 608 250 kr |
| Reduced Spillage (3%) | 3.00 % | 4 320 000 kr | 0 kr | 4 488 250 kr |
| Improved FCR (1%) | 1.00 % | 1 800 000 kr | 1 800 000 kr | 3 768 250 kr |
| Improved FCR (4%) | 1.00 % | 1 440 000 kr | 5 760 000 kr | 7 368 250 kr |
| Combined Improvements | 3.00 % | 4 320 000 kr | 5 760 000 kr | 10 248 250 kr |
AquaAI builds scalable AI, infrastructure and vision-based systems for aquaculture. We focus on robust pellet detection, appetite prediction and automated feeding support for both RAS and sea-based sites.
AquaAI operates independently from Dynamic Engineering AS and collaborates closely with Dynamic Fisheye AS on detection systems, feed control integration and applied research.
Based in Sørlandsparken, Kristiansand — reach out to discuss how AquaAI can help your farm.
Address: Stemmane 6, 4636 Kristiansand, Norway
Located in Sørlandsparken, Kristiansand — easily accessible from E18 and Kjevik Airport. We welcome on-site visits and live demonstrations.