FarmEye is in field trial on our Jaffna farm, and our farm apps run the team every day. We are also developing underwater inspection, cleaning and harvesting systems, with each project’s progress clearly shown.
Skip the 3D tour: go to the solutions ↓GMT FarmEye finds people coming from the shore or through the water and rings your phone with a siren. Nobody has to watch the monitor. In field testing on our own farm in Jaffna.
One films nets and stock in murky water. The other scrubs fouled nets while they stay in place. We publish results only after field trials.
A crawler on the pen floor finds each sea cucumber with a camera, lifts it with a soft gripper and keeps the harvest in a basket underwater.
Every batch boiled to the same recipe, counted and weighed, with a record you can trace. We test it with real batches before anyone relies on it.
Alerts, inspections and batch records in one place, with training, documents and support for your team. Start with one problem. If buying is better than building, we will say so.
Each comes with evidence before you commit. You always know which of the three you are paying for.
Cameras, AI detection and phone alerts, installed and supported by our team.
Put the machines we are developing to work on your pens as they reach each gate.
A paid assessment of one task on your farm, then a bounded build and support.
We run a sea cucumber farm in Jaffna, so these are our problems too. Pick a job to see what we are building for it, and how we will know it works.
A camera survey of your landing points, then a FarmEye pilot install.
| Today | Guards on foot or by boat, all night |
|---|---|
| Our approach | Long-range cameras, AI detection on the farm, siren and phone alerts |
| Benefit to verify | Intrusions caught while they happen |
| We measure | Alert-to-response time; false alerts per night |
A supervised inspection trial on your pens with the vehicle we are building.
| Today | Wading or diving to check nets |
|---|---|
| Our approach | Tethered underwater vehicle with camera and lights |
| Benefit to verify | Damage found earlier, less time in the water |
| We measure | Minutes per pen; damage found against missed |
Fund or host tank and pen tests of the net-cleaning crawler.
| Today | Scrubbing nets panel by panel from a boat |
|---|---|
| Our approach | A crawler with soft brush drums, driven from the walkway |
| Benefit to verify | Cleaner nets with fewer hours in the water |
| We measure | Net cleaned per hour; mesh wear |
Host tank tests of the soft gripper with live sea cucumbers.
| Today | Divers collect each sea cucumber by hand |
|---|---|
| Our approach | A seabed crawler with a camera and soft gripper |
| Benefit to verify | Less diving and gentler handling |
| We measure | Animals per hour; injury rate |
Share your process and volumes. We design the line around your recipe.
| Today | Washing, boiling and weighing by hand |
|---|---|
| Our approach | Washing drum, timed boiling tank and grading camera |
| Benefit to verify | Consistent batches with a record |
| We measure | Batch consistency; minutes per batch |
A camera survey of your landing points, then a FarmEye pilot install.
| Today | Guards on foot or by boat, all night |
|---|---|
| Our approach | Long-range cameras, AI detection on the farm, siren and phone alerts |
| Benefit to verify | Intrusions caught while they happen |
| We measure | Alert-to-response time; false alerts per night |
GMT FarmEye is built to catch people: poachers walking in from the shore or coming through the water. The AI watches the cameras, so nobody has to sit at the monitor, and rings the owner’s phone with a siren only when it finds a person or a boat. Built by SafeNet on Hikvision IP cameras for the pens of Globe Multi Traders (GMT) in Jaffna.
Field trialSNS-01 · 3D MODELOne camera recording, detecting and sending the first phone alert.
Control-room dashboard and the mobile app for owner and guards.
AI check of each snapshot, siren and PTZ automation, rules editor.
Farm computer, camera network and a second recorder.
Guard shifts, WhatsApp and voice escalation, more sites.
Real alerts from the field trial on our farm, captured from the owner’s phone. People are blurred. Not every alert is a poacher: the trial also counts false alarms.
GMT Worker for the team, GMT Accounts for the office. Workers check in with the farm’s QR code, the app counts their hours, tasks ring their phones, and wages, cash and books stay in one place.
A dated record for each project: what worked, what failed, what we measured and the next test. We add to it as we test.
Night detection at the pens: the AI boxed a boat beside the watch huts (73%).
A person alert from the Poompukar landing reached the owner’s phone at 5:22 am, with the snapshot.
False alarm: the AI marked a pen pole as a person (58%).
Detection range, false alarms per night and alert time are being measured; results are published after the trial.
Night range tests at each landing point, tuned zones to cut false alarms, and alerts sent after a link outage.
QR check-in, working hours, tasks, voice reports and payroll in daily use on our farm (version 1.6). Both apps are on Google Play.
Fingerprint check-in failed on phones without a strong fingerprint sensor. Fixed with a new device-key method.
P1 engineering package: 100-part CAD, 14 drawing pages and a 21-page build guide.
Control software: 65 tests pass on the bench, in dry run (no motors connected yet).
No physical prototype and no water test yet.
Build P1, bench-test power and fault shutdown, pressure-check the housings, then a controlled-water test.
No prototype or field record yet. The first dated test results will appear here.
These are in development or at concept stage. The 3D models and films show the intended design, not a tested product. Hover a card to see inside. Open one for its parts, film and test plan.
In developmentSNS-02AquaSentinel inspects nets, poles and stock in murky lagoon water. The P1 design is complete; the first build is next.
Next gateBuild P1, then pressure and controlled-water tests
ResearchSNS-03A crawler that scrubs fouled pen nets so water keeps flowing through the mesh.
Next gateTank test on a net panel: grip, coverage and mesh wear
ConceptSNS-04A seabed crawler with a camera and a soft gripper that collects live sea cucumbers. No divers.
Next gateTank test: find, grip and lift live sea cucumbers without damage
PlannedSNS-05Washing, timed boiling and grading in one line, with a record for every batch.
Next gateTest batches: boiling consistency, piece damage, weighing accuracy
The assessment decides what is feasible. Then we build the smallest thing that solves the measured problem.
Water quality, power, pumps and aerators, with alerts to the right person.
Count, size or check stock and product, measured under your lighting and water.
Add sensors, controls or drives to boats, pumps and equipment you already own.
Harvest, batch and export records in one place, ready for buyers.
Software servicesVideo, levels, power, counts.
Runs detection and control. Works without internet.
Phone alerts, history and reports for the team.
Added only when the data supports it.
A short call or a video of the task: what, how often, the output, the issue and the location.
We observe the baseline and compare three options: buy equipment, adapt what you have, or build.
A rig, software proof or modified part that resolves one defined uncertainty.
It runs in your real conditions while we record performance.
Install, commission, train your team and hand over documents.
Assessment fee = engineering time + travel + measurementsFixed before the visit.
Pilot price = design + parts + fabrication + controls & software + assembly & testing + contingencyQuoted after the assessment, with what is and is not included.
Monthly items = cloud + connectivity + serviceAlways quoted separately, never hidden in the build price.
Payments = enough upfront for committed parts, then milestones you can reviewChanges go through written change control.
Count losses avoided, extra energy, maintenance, downtime and support in the benefit. Staff time saved becomes a cash saving only if spending actually falls.
SafeNet Solutions is the engineering and automation division of SafeNet Creations, based in Jaffna. We start with aquaculture, because we farm sea cucumbers ourselves and can test every idea where the problems are real.
Our team and collaborators include engineering graduates and student contributors with academic backgrounds at the University of Moratuwa and the University of Jaffna, alongside SafeNet’s software and AI team. Together we investigate farm problems, develop designs and turn suitable ideas into working systems.
Customers work through SafeNet for project coordination, clear milestones, training and agreed support.
One accountable person for scope, schedule and your questions.
Experienced engineer who signs off designs and acceptance tests.
Wiring, protection, controllers and safe local control.
Local workshops for machining, welding and enclosures.
SafeNet’s developers for dashboards, alerts, records and vision.
Research, CAD, coding and test documents, under supervision.
Prefer to write it down? Describe your farm problem