AI health monitoring for fish farms, hatcheries & shrimp ponds.
AquaHealth is B2B SaaS for commercial aquaculture operators, starting with fish and shrimp farms in Nigeria. Upload the inspection photos your team already takes — from tanks, ponds, cages, or raceways — and our vision model flags early signs of disease, stress, and abnormal behavior before they show up in your mortality numbers.
No cameras to install. No sensors to wire up. One subscription covers every site you run.
Fin erosion pattern detected · 82% confidence · Tank 4B
We're early. AquaHealth is a working product in active pilots, not a finished platform — every flag is meant to be reviewed by your team, and we're upfront about what the model does and doesn't cover yet.
Visual inspection doesn't scale past a handful of sites.
Most disease and stress indicators show up visually first — before water chemistry moves, before mortality spikes. The catch is that spotting them depends on a trained eye walking every tank, pond, or cage, every day, and remembering what changed since yesterday.
Once a fish or shrimp operation grows past a couple of sites, that just isn't consistent. Photos already get taken on rounds — we built AquaHealth, a B2B SaaS platform, to make sure they get a second, tireless set of eyes.
One platform for every kind of fishery operation.
AquaHealth is purpose-built for commercial aquaculture — not general animal health or consumer wellness. If your team walks tanks, ponds, cages, or raceways for a living, this is for you.
Finfish farms
Salmon, tilapia, catfish, and other finfish operations running net pens, cages, or ponds.
Shrimp & crustacean farms
Pond-based and intensive shrimp producers screening for stress and disease across grow-out cycles.
Hatcheries & nurseries
Early-stage rearing operations where catching a problem a day sooner protects an entire cohort.
RAS & land-based systems
Recirculating and indoor operations that need consistent monitoring across many small tanks.
Three steps, no new hardware.
A cloud dashboard your whole team logs into — subscribe once, and it covers every tank, pond, or cage across every site. Nothing to bolt on and nothing to wire up before your first flag.
1. Upload from any phone
Your team takes the same rounds they already do and uploads photos through our web dashboard — no proprietary camera, no install visit.
2. The model screens each photo
A vision model trained on labeled aquaculture inspection images checks for known lesion, parasite, and stress-behavior patterns and scores its confidence.
3. Your team reviews and confirms
Flags land in a shared queue, not an autopilot alarm. Confirming or correcting a flag is also how the model gets sharper for your site.
See what a flagged photo looks like.
A scripted walkthrough of the review experience using illustrated example scenarios — not a live upload yet. Book a call and we'll run it on your real photos.
Pick a scenario on the left, or hit “Routine pond check” again to run it — this is a scripted walkthrough of what your review dashboard looks like, not live inference.
Book a walkthrough and we'll screen a few of your real inspection photos live.
Built and running — still early, and we say so.
Working product
The upload-and-screen pipeline runs today. Model coverage is growing condition by condition, and we tell you where confidence is low.
Onboarding pilot farms
We're taking on a small number of design-partner operators to shape the product with real inspection photos, not just our own test set.
Human in the loop, on purpose
Every flag is a suggestion for your team to confirm, not an automated verdict. That's deliberate while the model is still young.
A small, technical team building AquaHealth.
Two co-founders, one product: getting a useful vision model in front of the people who walk the tanks every day.
Read our full story
Onyedika Onuorah Ikechukwu
Leads product and the early pilots — spends most weeks talking to the operators actually using the tool.
Peace Oyaleke
Machine-learning engineer. Builds and trains the vision models behind every flagged photo.
