My garage door opener is one of the most boring pieces of tech in my house, made by a company that’s 67 years old. The one I bought last year looks almost identical to the ones it released a decade ago. But for the past 16 months, a small band of developers at Chamberlain has been quietly working on a gadget like nothing else on the market.

They’ve built a fully-automated, internet-connected pet door for your dog, one that gives them the freedom to go outside whenever they want, or lets you remotely activate it yourself. Your pet can approach the door to ask you for permission while you monitor its cameras from an app — one that’ll also automatically track your pet’s comings and goings in a diary.

It’s called the myQ Pet Portal, and it’s packed with tech, including twin 1080p cameras that beam encrypted video to your phone anywhere in the world via Amazon S3 servers, IR and light-touch safety sensors to avoid pinched tails, microphones and speakers to talk to your pet, plus the encrypted Bluetooth Low Energy beacon that goes around your dog’s neck. The door won’t open until your pet is stationary for a few seconds to make sure they actually want to go out, and it closes automatically behind them.

Preorders for the myQ Pet Portal start today at an eye-watering $3,000 and up, which at first blush makes it seem like exactly the kind of ostentatious, far-out product concept that graces the CES show floor in Vegas every year — an awesome idea to gawk at, but not something anyone actually needs. After all, can’t you already buy a simple pet flap for well under $100?

But the world’s largest garage door opener company seems dead serious about building the best doggie door ever made. During the COVID-19 pandemic, roughly a dozen staffers have been toiling away in their own homes — guided by extensive surveys, focus groups and testimonials about what pet owners might need to actually trust a product that lets their furry companions in and out of the house.

Today, says product lead Beril Altiner, pet doors don’t necessarily have that trust. Only 34 percent of pet owners who let their pets outside have a pet door, according to a national survey conducted by the company. That means 68 percent are doing without — forcing pets to wait for humans to let them outside, forcing themselves to schedule their lives around their pets, or both. “There’s a lot of guilt among pet parents in their ability to tend to pets’ needs in a timely fashion,” says Altiner, showing me both survey results and video testimonials from individual pet owners.

The team thinks that’s partially because most of today’s pet doors involve making a visible hole in your house. Some make you actually cut a hole in your door. Others slide into your sliding glass door track, generally requiring you to add weather sealing to keep the breeze out. And all but a handful involve a flap light enough for your dog to push it out of the way, potentially allowing the cold or stray animals to get in. Many pet owners that do buy pet doors wind up abandoning them, says Altiner. (I have some personal experience there.)

So instead of building yet another flap, the garage door opener company set out to build an entire door of its own — with a doggie portal that would completely seal and lock when it closes. Preferably, it’d be a pet door you wouldn’t even know was there.

An early concept video shows off what that might look like, and Chamberlain actually prototyped a rough working version last January, too.

At this point, they knew they could build a motorized, solid, weather resistant barrier that’d meet building codes and UL requirements… but in order to create an opening big enough for large dogs, they found, the stealthy vertically closing door wouldn’t have enough room left for a window. By June, an additional survey of over 500 homeowners showed the team going windowless wouldn’t be wise: 29 percent of hinged doors used by pets are full glass, and 76 percent of doors dedicate at least a quarter of their height to a window.

Some respondents said they’d give up their glass, but enough people valued the light that the vertically sliding concept had to be shelved. “They are too big for us to ignore as a target,” reads one of Altiner’s notes in an internal presentation.

The next phase was low-tech: the team cut holes in cardboard boxes and ushered some dogs through. Chamberlain still had five different mechanical openings on the table that left room for a window, but since survey respondents didn’t seem to have a preference, they used cardboard cutouts to test with actual dogs. A mechanically lifting flap left room for a 125-pound dog, it turned out, but it’d be more complicated and less safe than the alternative they eventually picked: a pair of elevator doors that could use the same infrared safety sensors that Chamberlain invented in the ‘80s and uses in all its garage door openers today. The design meant they’d have to settle for up-to-90-pound dogs, but they were now ready to build.

By now, well into the pandemic, the Chamberlain team had to send out jury-rigged test kits to its firmware and app developers, consisting of a board with only the raw sensors, motors and switches they’d need to manipulate, because those developers would rarely get to see a real door. In fact, it wasn’t until July that the company found a door-making partner — Kolbe — and September until they tried to put one inside an actual slab. “We worked with our partner for the longest time without ever meeting them in person,” Altiner says, recalling one day she had to wake up at 4:30 in the morning, drive four hours from Chicago to Kolbe’s facilities in Wausau, Wisconsin… then back again at 10:30 in the evening because COVID restrictions meant she couldn’t stay in a hotel overnight.

The team had a handful of other in-person, socially distanced meetings for big tests, but it was largely Slack, Microsoft Teams, and a whole lot of video calls and emails to put it together.

Things didn’t always go smoothly. In September, the first real door failed its Air, Water and Pressure test, leaking liquid into the wooden slab. A Very Good Girl bumped into the too-high step in an attempt to get through. The door’s differential-driven gears churned like an overzealous egg beater. The team had to change the entire way it sealed the module to keep the water out, and switched the noisy gearbox out for a quieter direct-drive for the sequel.

The second set of doors (Alpha 2), the first to hang in actual testers’ homes, didn’t fit properly when they arrived in October — apparently the distant teams weren’t quite on the same page about measurements. They also had conspicuously dangling power cables, because they hadn’t yet routed power across the door’s hinge itself. But by December, the third set passed all the tests, and last week Chamberlain was ready to show me — via live video chat, of course — how it actually works in the field.

This is the part where I tell you I haven’t seen this door in person, so I can’t truly say, but I didn’t get the sense there was anything to hide. There are still loads of seams that make it pretty clear there’s a dog door in there — RIP, vertical sliding panel — but it looks like a real door with a big window, the elevator doors slide right open, and it seems nice and quiet. Chamberlain product manager Greg Martel had to point his phone right up close before I could hear the motors engage and the doors slide along their tracks. Fenway, a 9-year-old, 45-pound Lab Collie Spaniel mix, seemed to have no trouble opening it by herself.

The app doesn’t do a lot right now, though. Like Chamberlain’s myQ garage door openers, the app shows you how long the door has been open, and the video and audio streams work, but tracking your pet’s outings, video recordings, even the notifications that let you know when your dog wants to go out are still to come. Here’s what the company wants the app to look like when it’s done:

And as you might expect from what I’ve described of Chamberlain’s CES sprint so far, it’s not the only thing the team needs to finish before it starts shipping to the first batch of buyers, tentatively around April. Last I checked, they were still figuring out whether it’d be rolling out nationwide or just regionally, because it wants to hire a set of trusted installers to put this door in your home. “This is going to be a pro installation only,” program leader David Schuda tells me, partially since the Pet Portal requires power through the hinge, but also because you can opt for a pre-hung door that comes with a doorframe. Installation isn’t included in the price, by the way.

Speaking of pricing, it’s only recently that the company finalized how much you’ll pay — originally, Chamberlain told me the starting price would be closer to $1,800, before raising that 66 percent to the $3,000 it’s asking now. That’s because originally, the entry-level door would be an unpainted, windowless model, something the team changed their minds about when they remembered how few homeowners had solid doors leading to their yards.

For now, the fiberglass door will start with a 36 x 21-inch pane of glass, with a slightly larger one in the pricier wood model. And while $3,000 is pricier than anything else on the market, there’s nothing else that quite compares. The company justifies it saying it’s half the price some pay a dog walker each year, which — at 252 business days per year, assuming one walk per day at $20 each — seems like it could be true.

There’s a pretty simple reason I won’t be buying a myQ Pet Portal, though, even though I love smart home gadgets and have a dog whose schedule often doesn’t meet our own: I’m in the 16 percent of doggy homeowners with sliding glass doors instead of swinging ones, a group Altiner and co admit they’re not going to be able to serve right now. But I came away fairly convinced this is both a good idea and not a piece of vaporware, which is pretty remarkable for a one-of-a-kind, multi-thousand-dollar CES product announcement.

Just to cover my bases, though, I asked Chamberlain president and COO Jeff Meredith to commit to that on the record. “I want the notoriety of the product, but I also need the revenue of the product,” he told me. “It’s going to ship.”

If this works out, I’m curious what Chamberlain might do next. The garage door opener company might not be the most exciting tech firm in the world, but it’s had its moments, like roughly a decade ago when it introduced the first one you could open from anywhere in the world, and more recently when it partnered with Amazon to deliver packages to your garage.

It might be lip service, but Schuda tells me that over 20 years with the company, he’s never seen Chamberlain spend as much time researching what people might actually want and need when building new products. Before, “we had featureitis, we’d just add features that we came up with,” says Altiner. They credit Meredith for pushing the research-driven approach, something that might sound familiar if you’re a long-time Verge reader — we wrote about his similar initiative at Lenovo in 2016, though that idea didn’t necessarily work out.

“This is technology aimed at solving real people’s problems,” says Meredith, of the kinds of things he now wants Chamberlain to build. How many real people’s problems can you solve at $3,000? We’ll see.