01 Sense
One handheld suite reads dissolved oxygen, pH, TDS and temperature across every pond, then returns to a solar dock to charge.
We automate high-density fish and shrimp farming
One low-cost sensor suite covers the whole farm. Our AI diagnoses faults and tells farmers what to do over WhatsApp.
The whole farm runs on one sensor suite
Keep scrolling
The handheld monitoring unit
Dissolved oxygen, pH, total dissolved solids and temperature are read across every pond on the farm, and the unit returns to a solar dock to charge. One suite costs a fraction of a fixed sensor array, which is the reason a smallholder can afford one at all.
See it in action
Automation in three simple steps
02 Diagnose
The inference engine ranks likely fault causes from field data, so a drifting number becomes a clear explanation instead of a guess.
03 Act
Farmers get the next action in WhatsApp, in their own language. Owners oversee production from the dashboard.
A closed feedback loop improves the farm every cycle
Keep scrolling
An AI agent handles every conversation with the farmer
Operators never touch a dashboard. The agent sends each instruction in their own language through WhatsApp and takes their readings back the same way, which keeps the field workflow as simple as a chat. Owners get the dashboard instead, where they oversee production across ponds and analyse the data behind it.
One unit covers every pond on the site
The unit moves between ponds rather than sitting in one, so a single suite reports the whole farm through the same agent. Adding a pond does not mean adding hardware, which is what makes cost per pond fall as a farm grows.
From the pilot
One commercial cycle produced this
Sixteen weeks of Nile tilapia at the pilot, against a conventional flow-through pond over the same period.
47% less feed for the same weight of fish
Feed conversion of 0.846 against a typical flow-through 1.6. Feed is the largest running cost on a fish farm.
25 kg/m³ stock density
held for a full cycle
Ten times the fish in the same water versus roughly 2.5 kg/m³ traditional, at 84.89% survival.
Biofloc holds ten times the fish in the same footprint
Microbes convert fish waste back into protein in the water column. Ammonia stays down while stocking density goes up and water use drops. The biology has been proven for years.
- HDPE-lined tanks run continuous active aeration
- Floc recycles waste nitrogen into feed in the water column
- Water exchange is a fraction of a conventional pond
- One missed overnight oxygen crash still costs the whole harvest
High density has always demanded an expert on site
That labour constraint keeps intensive farming out of reach for independent producers. Software removes it.
United States
Consumption rises every year and between 75 and 90 percent of it is imported. Domestic producers cannot reach intensive density because the labour to run a farm at that level does not pencil out.
Indonesia
There are 6.5 million ponds and three quarters of them run at low density. The infrastructure is already built and already paid for, which makes Indonesia the fastest place in the world to add density.
The pilot beat its target growth curve
A full commercial cycle was tracked week by week against the target curve for Nile tilapia. Realised biomass finished ahead of target at a feed conversion ratio below 1.0.
This is the farm the numbers came from
Footage taken during the commercial cycle, from stocking through daily feeding to the sampling that produced the growth curve above.
The system runs on working farms today
We work with farmer cooperatives and independent producers across West Java in Indonesia and in Escondido, California, running the full stack from monitoring on a solar dock through the inference engine to the operator dashboard. We are looking for new farms to work with.
Research collaboration
Expert input is built into the product
Floc.ai runs an international cross-sectoral research collaboration spanning three countries. Funded student engineering and aquaculture science feed directly into the knowledge base, so every new farm and every new institution makes the model better for every farm already on the network.
Worcester Polytechnic InstituteInference model development
IPB University, BogorAquaculture biology and experimental ponds
Jakarta State PolytechnicSensing hardware and field deployment
BBPAT
Center for Freshwater FisheriesFarmer network access
Every institution adds engineering capacity and field data at the same time, so the model improves for every farm on the network. That compounds.
A number that drifts is not a fault. Readings pass through a fuzzifier and a knowledge base built from field data, and the engine ranks the likely causes across its fault classes before deciding whether the answer is guidance to the farmer or a direct change to the aerators.
Team
Founders
CEO and Co-founder
Kemal Rifky
Mechanical Engineering master's from WPI and a second-time founder who built and sold a solar energy startup in Indonesia. Leads aquaculture science, hardware and field deployment.
Kemal Rifky on LinkedIn
CTO and Co-founder
Ben Tyler
MSc in Artificial Intelligence at Queen Mary University of London with published human-computer interaction research at ACM CHI. Leads engineering, machine learning and product.
Ben Tyler on LinkedInSee the system running
A walkthrough of the full stack at a partner farm takes you from sensing to diagnosis to the instruction landing on a farmer's phone.
Player not loading? You are probably opening this file directly. Watch on YouTube
Ten times the density needs no expert on site
We are building the intelligence layer for the next generation of fish and shrimp farms.