Virtual fencing · early-stage prototype
A cattle fence that fits in an ear tag.
HerdTag is a small GPS ear tag that keeps cattle inside a pasture you draw on your phone. It warns with sound when an animal walks toward the line. No collar, no new posts, and no shock built into the tag.
The problem
Moving fence and moving cattle eat days.
Fencing is labor
Rotational grazing pays off, but every new paddock means posts, wire and hours. Cross-fences are hard to justify on rented ground or land that changes use each season.
Collars are heavy
Today's virtual-fencing systems mostly use neck collars. They work, but they have to be fitted to each animal, adjusted as it grows, and paid for.
Ear tags are already normal
Every producer already tags cattle with a standard applicator. If a tag could do the fencing, adoption could be as simple as tagging day.
How it works
The tag decides on its own. The network just keeps you informed.
Each tag carries the pasture boundary in memory, so the fence keeps working even when the farm's internet is down. The tag sends a short check-in every 15 minutes, and sends one right away when something needs your attention.
- Ear tagGPS, fence logic and sound, all on the animal
- Farm gatewayLoRa radio with miles of range, no cell coverage needed
- HerdTag serverhistory, alerts and fences
- Your phonedraw pastures and get alerts
Draw
Outline a pasture on a satellite map and send it to the herd.
Watch
Each tag checks its GPS more often the closer the animal gets to the line.
Warn
An animal walking toward the line hears a short sound. It stops as soon as she turns.
Alert
If an animal gets out anyway, the tag stops cueing and you get a notice right away.
Progress
What exists today
The software and the build plan are ahead of the hardware on purpose. Everything that can be tested on a laptop gets tested there first.
Fence controller
The on-tag logic: warning zone, cue only when moving toward the line, never cue on a bad GPS fix, hard limits on cues.
Herd simulator
Simulated cows with realistic GPS error, used to test the logic and estimate battery life before touching an animal.
Firmware core
A C version of the controller for the tag's microchip, checked against the reference version on thousands of test cases.
Server and map
Takes in tag messages, stores history, lets you draw fences and send them to tags, and tracks which tags have confirmed.
Dashboards
A plain-language farmer view, an operations dashboard and a trial view for the vet and behavior specialist.
Hardware prototypes
Ten tags built from off-the-shelf modules at a makerspace, in 3D-printed cases, tested on people before cattle.
Roadmap
Each phase has a pass/fail checkpoint.
We run the cheapest experiment that could prove us wrong first. If a checkpoint fails, we stop or change course instead of building on hope.
- Phase 0
Desk checks
Talk to producers and researchers already working on this. Check the patent landscape. Confirm a working farm can host trials.
Pass if: the farm fits, nothing blocks us legally, and we're not duplicating work already done.
- Phase 1
Bench prototype
Build ten tags and a farm gateway. People carry the tags around a marked-out pasture to test GPS, radio range, cue timing and battery.
Pass if: at least 95% of messages arrive, cues fire at the right moment, and the power budget works with solar.
- Phase 2
Cattle trials
A small, supervised trial inside an existing physical fence, with a vet and a behavior specialist. Pass/fail thresholds are written down before the trial starts.
Pass if: cattle learn to turn at the virtual line, at an escape rate the farmer would accept, with no welfare red flags.
- Phase 3
Moving cattle
Only if Phase 2 passes: try guiding a trained animal, then a small group, along a route to a gate, comparing one tag against one in each ear.
Pass if: animals follow guidance consistently and calmly.
Animal welfare
Rules we hold ourselves to
- No shock in the tag. The tag cues with sound, and possibly vibration later. Nothing else.
- Cues are capped. There are hard limits per episode and per hour. If an animal ignores a cue, the tag stops and reports it. It never gets louder.
- Out means quiet. Once an animal is outside the line, the tag stops cueing so it never discourages her from coming back.
- No cue on a bad fix. When GPS is unsure where the animal is, the tag stays silent.
- Vet-supervised trials. A veterinarian fits the tags, checks ears daily, and can stop the trial at any time.
- A real fence is always there. Early trials run inside existing physical fencing. A virtual fence is never the only barrier between cattle and a road.
Questions
FAQ
Can I buy one?
No. HerdTag is a research prototype. If the cattle trials go well we'll say so here and share the results, including if they don't.
Does it work without cell coverage?
Yes. Tags talk to a LoRa gateway on the farm, which reaches miles in open country. The fence logic runs on the tag itself, so a dropped connection doesn't turn the fence off.
How long does the battery last?
We don't know yet. Battery life depends on how much time animals spend near the boundary, because that's when the tag checks its GPS most often. Our model gives anywhere from days to months depending on assumptions. Measuring it is part of Phase 1. Try the estimate yourself.
Will it replace my official ID tag?
No. HerdTag sits alongside the animal's official identification tag and doesn't replace it.
Why not just use sound plus a shock, like some collars?
Those systems have good welfare research behind them. But a shock in an ear tag is a different and less-studied thing, and many producers would prefer to avoid it. We want to learn what's possible without one first.
Can it move cattle between pastures?
That's the long-term goal and the hardest part. We won't attempt it until containment is proven.
Collaborate
Ranchers, researchers and makers welcome.
If you run cattle and want to help shape this, study livestock behavior or welfare, or build low-power hardware, we'd like to hear from you.