Varroa control that doesn't come out of a packet.

Spot on — optics can't see through wax. But a mite has to get into a cell before it's capped, and it has to come back out when the bee emerges. Both of those happen in the open, and both run to a schedule we can read from scanning the same frame every day.

So we clear cells in the day or two before they're capped, and we're watching when they open. Fewer mites get sealed in; fewer walk away unchallenged at emergence. The full explanation is here.

It isn't perfect and we don't pretend it is. What it does is keep pressure on the mite population every day — and on the other parasites we add over time — rather than four times a season — and heavy loads will still want a brood break and an oxalic dribble.

Because a flat "no resistance, ever" would be a claim about evolution that nobody can back, and you'd be right to distrust it.

There's no known biochemical route for a mite to become resistant to being physically destroyed. That's genuinely different from a chemical that binds to a receptor — with amitraz or flumethrin a single point mutation is enough, which is exactly what's happened here: L925I for the pyrethroids, Y215H for amitraz. Nothing equivalent exists for a photon.

But mites are under selection pressure from anything that kills them, and behaviour is a plausible route. Mites that spend less time on adult bees, or avoid a lit transit path, or ride in positions that are harder to see, would be selected for. Nobody has tested that, because nobody has run a device like this across enough seasons to find out.

So: no known mechanism, far narrower than chemistry, and not zero. We'd rather say that now than walk back an absolute in five years.

Combplex is real and doing good work. They've been at laser varroa control since 2018, announced pre-sales of their first 100 frames in February 2026, and got here before us. We're not going to pretend otherwise.

The difference is what the device does. Theirs is a fixed detector array in the corner of a frame — bees walk past it, and mites riding on them get hit. Ours is an X–Y positioner that scans the comb face itself, which is what lets us work the brood cycle: clearing cells in the day or two before they're capped, and being there when they open. That's set out here.

Beyond the mechanism: a different form factor, other frame depths on the way, and a pricing model we haven't fixed yet. And most commercial beekeepers already run multi-vendor stacks — two laser companies in a market this size isn't crazy.

v1 is varroa only, and we'd rather be clear about that than sell you a roadmap. But the mechanism isn't varroa-specific — a camera, an X–Y positioner and a laser is a general way to find a small foreign object on a bee and act on it. Small hive beetle is the obvious second target and optically it's an easier problem than varroa: several millimetres rather than one and a half, darker, and hard-shelled against pale wax. Removing an armoured beetle that runs is the harder half, and we haven't solved it.

The full list of what we think this can and can't ever reach — including the things it will never touch, like nosema, foulbrood and tracheal mite — is set out here.

It's the concern we take most seriously. We're commissioning an independent, peer-reviewed bee-safety study — sub-lethal effects, foraging, learning, brood viability, queen acceptance — before any commercial sale. If the science doesn't come back clean, we don't ship. Full stop.

On the hardware side, the design target is a fully enclosed IEC 60825 Class 1 product, which means no eye-safety obligations land on you as the operator. Anything above Class 1 would put radiation-safety licensing onto beekeepers, and that's not a product anyone should build.

Fair question, and the whole design point is that you don't fiddle. Drop it in at the start of the season, take it out at the end. Telemetry is on your phone if you want it. If you don't want a phone in this conversation at all, we'll send you a printed monthly report.

Yep. That's the engineering problem we spend most of our time on — optical surfaces, wire routing, heat dissipation, form factor. All of it has to survive a real apiary, not a lab bench. Which is exactly why we want a trial in your yard rather than in an office. We'd rather you tell us what's broken than find out eighteen months later.

We won't trash anyone else's product, but we'll say this: we're building a thing that does something. Sensors that only measure don't change your day. Thermal treatments are temperature-sensitive and can cook brood. The passive grooming category's one independent peer-reviewed trial found no significant effect. We're trying to actually remove the mite, without a chemical and without adding to your workload — and if we can't, we won't pretend otherwise.

There's a lot of hype out there. The vision component does what your own eye does when you watch a bee crawl across a frame — it sees a mite. The model's job is to tell a mite from a bit of pollen. It isn't flying the plane. The decision to remove a varroa is the same physical decision you make when you spot one. We just do it many thousands of times a day, in one hive, without getting tired.

Honest answer: hardware companies die more often than software companies do. Two protections. The hardware works without our cloud — detection and counting run on the device. And we intend to publish enough of the protocol that an installed unit could be maintained even if we vanished. We're not building a roach motel.

A physical control device that emits no substance shouldn't fall under the APVMA's agvet chemical registration regime — the Agvet Code is written around substances, and thermal varroa devices are already sold in Australia without registration. But we're seeking written confirmation from the APVMA rather than assuming, and we'll publish what they say. Varroa remains notifiable in every state and nothing about this device changes your reporting obligations.

We're not taking orders and we won't until there's independent efficacy data. The realistic sequence is: engineered prototype, supervised trials across two seasons with named commercial beekeepers, a published study, and then a product. If you'd like to be part of the trials, the form above is the fastest route.