We are building automated hydroponic growing racks for small and vertical spaces, each one specialised for the crop it grows. The rack keeps water, nutrients, light and climate right, day and night, with AI running on site that is built to learn from every grow - from a single room today to industrial farms tomorrow.
Growing indoors makes harvests local and predictable. What holds it back is not the idea - it is how hard an indoor farm is to run.
In February 2023, UK supermarkets rationed tomatoes and cucumbers after bad weather hit growers in Spain and North Africa. A shelf can empty because of a cold week somewhere else.
A closed room near the people who eat the food: no seasons, no droughts, a fraction of the water, harvests planned by the calendar instead of the forecast.
In a hydroponic tank the water never stands still: pH and nutrients drift, oxygen falls, temperature climbs. Left alone for a few days, the solution goes bad. Today that means someone checking, every day.
Three layers designed together, so the farm looks after itself and the grower decides what to grow, not when to top up a tank.
A modular rack of stacked growing zones, built for small and vertical spaces. Each zone is specialised for its crop - its own light, its own climate, its own nutrient loop - so lettuce, basil and strawberries can grow side by side, each on its own program.
Small smart devices, each with one job, all connected through a gateway. Plug a module in and it introduces itself to the system: the farm grows one module at a time.
GROWMATIC® STUDIO is more than an app: an integrated development environment for agriculture. Program your grow phase by phase and the system follows it; every dose and every correction is logged - a grow you can repeat, audit and improve.
Everything in the system serves one loop: notice the drift early, correct it in small steps, check that the correction actually happened - and write it all down.
The system reads the water and the air of every zone, day and night.
It responds in small steps - only what is needed, only when it is needed.
Each module reports what it really delivered. No dosing in the blind.
Every action is logged, and AI turns the data into a better next grow.
Every grow produces data: each dose, each reading, each harvest. We already run AI models locally, on our own hardware - no cloud dependency, no data leaving the site. Next, we adapt those models to growing, so the system does not just follow a program: it learns to write a better one.
AI models running on our own hardware. Data stays on site and nothing depends on someone else's cloud.
Models adapted to our crops and our data: every logged grow makes the next one smarter.
Questions in plain language: why did pH drift last night? What should change for the strawberries? Answers from your own data.
Suggestions on light, nutrients and climate, refined grow after grow - and the grower always has the last word.
Growing bigger should not mean starting again. The control layer - protocol, programs, software - stays identical at every size; only the pumps and valves grow.
A single rack in a spare room, a shop or a kitchen. The place to start, and to learn.
A sealed, portable farm: racks inside, natural energy outside. Grow wherever there is space.
Zones stacked floor to ceiling, each on its own crop, all driven by one gateway.
The same modules in parallel, heavy-duty pumps on the same bus. Our longer-term goal.
Next step for our racks: leaving the workshop and growing for real. We are looking for a place to host them for 6 to 12 months, in Greater London or the Home Counties.
Most automation in this space only works with its own parts. We are going the other way: every module speaks Modbus RTU, an open industrial standard, and we will publish the full register map so anyone can build a compatible device or plug ours into an existing setup.
Standard Modbus RTU over RS-485, end to end.
Register map and self-declaration spec, documented in the open.
Use one module or the whole system - your choice.
We are at an early stage, and we would rather show the work than promise a launch date. These are the goals in front of us.
Complete the first rack build: frame, removable trays, lighting and climate, with GROWMATIC STUDIO running the loop.
Find the location where the racks are deployed and run for real - real crops, real drift, months instead of hours.
Turn the prototype into a rack that can be built again and again: structure, trays, water paths and assembly engineered for repeatable production.
Automate the hands-on work too - moving trays, seeding, harvesting - so a rack needs a visit, not a shift.
Deploy a 20ft container to experiment with natural energy sources powering a complete growing system.
Adapt the AI models we already run locally to our crops and our data.
Heavy-duty versions of the same racks and modules, and a first industrial pilot.
We want the system to take shape around the people who will use it. If any of this sounds like you, we would love to hear from you.
Growing at home, in a container or on a farm? Follow the project and help shape what we build first.
Follow the projectRunning an indoor farm, a research setup or a business that could use this? Let us talk about a pilot.
partners@itamatic.comInterested in where indoor agriculture is going and how we plan to get there? Write to us.
team@itamatic.comOccasional updates as the project comes together: the rack build, test results, the container farm. No spam, unsubscribe anytime.
Follow the project
See our post on LinkedIn 
Instagram