Daniel's been digging through the museum of comedic tech failures and landed on one of the strangest artifacts of the dot-com era. The CueCat. This was a barcode scanner shaped like a cat, given away for free by RadioShack around the year two thousand, and Daniel wants to know the whole story. What was it, how did it work, what was the grand plan to put one in every home, and why did it crash so spectacularly? And the practical question — if you found one in a drawer today, could you still make it do anything?
The answer to that last one is yes, with some surgery and some very specific adapters. But the story of how we got there is...
Let's start with the object itself. Because the hardware tells you almost everything about the ambition and the flaw.
So picture this. It's the year two thousand. You walk into a RadioShack and there's a cardboard display with a cat-shaped cutout, and inside are these translucent blue or purple plastic things. They're about the size of a computer mouse, but shaped like a cat. The tail is the handle. The nose is where the scanner lives. It connects to your computer through a PS/2 port — that round purple or green connector for keyboards and mice — and it sits between your keyboard and the computer. It's what's called a wedge device.
A keyboard wedge. Which is a term that somehow sounds both obscene and boring.
It means it literally sits in the middle of the keyboard cable and injects keystrokes. To the computer, the CueCat looks like a very fast typist. You scan a barcode in a magazine, and the CueCat decodes it, converts it to a string of text — usually a URL — and then types that string into whatever application has focus. No drivers needed. No software installation, technically, though they did ship a browser plugin to handle the URL launching gracefully.
So the wedge architecture is the first thing that tells you about the ambition. They wanted zero friction. No USB drivers to install, no configuration, just plug it in between the keyboard and the tower and suddenly your magazines are clickable.
Right. And the company behind this was called Digital Convergence, founded by a guy named J. Michael Fitzpatrick. The pitch was ambitious. Every magazine ad, every catalog page, every product label would have a special barcode printed on it. You'd scan it with your CueCat, and your browser would open directly to the product page, or a related website, or some kind of interactive content. No typing URLs, no searching. Just point at paper and click.
And the business model was that advertisers would pay for the barcodes, and consumers would get the hardware free. RadioShack committed to distributing millions of units. Forbes and Wired ran articles about how this was going to merge print and digital. They raised over a hundred and eighty million dollars.
A hundred and eighty-five million, by most accounts. Advertisers like Pepsi, Procter and Gamble, and RadioShack itself signed on. The vision was that every home would have a CueCat sitting next to the computer, and people would just... scan things. Magazines, catalogs, maybe eventually product packaging at the grocery store.
And this is where we hit the first thing that most people get wrong about the CueCat. It wasn't a standard barcode scanner. It didn't read UPC codes.
That's the key. Digital Convergence invented their own proprietary barcode symbology. It was a dense, high-capacity two-dimensional barcode — it looked a bit like a QR code, but it was completely custom. And the CueCat was designed to only decode that format. So you couldn't take your free CueCat and scan the barcode on a box of cereal. It would just... sit there. It only worked with CueCat-branded content.
Which is the fatal lock-in, right there. They gave away the razor but only sold one brand of blade, and the blades were printed in magazines that you had to go find and hold up to a plastic cat.
And the proprietary format made a certain kind of sense from their perspective. If the CueCat could scan any UPC code, then what's the business model? They needed to control the barcodes so they could charge advertisers to create them. The barcode was the ad unit. The cat was just the delivery mechanism.
So you've got this cat-shaped wedge. But who thought this was a good idea, and how did they convince investors to pour a hundred and eighty-five million dollars into it?
The team had some genuine credibility. Fitzpatrick had connections. Several former RadioShack executives were involved, which is how they got the distribution deal. And you have to remember the context. This is nineteen ninety-nine, two thousand. The dot-com bubble is inflating. Anything that promised to bridge the physical and digital worlds was getting funded. QR codes existed in Japan but weren't on anyone's radar in the US. The idea that you could point a device at a printed page and have your computer respond — that felt like the future.
And the wedge architecture was clever, in a vacuum. The CueCat decoded the barcode into a string, and that string got injected into the keyboard buffer. The CueCat software — a little helper app or browser plugin — would watch for a specific prefix pattern and then intercept the URL, launch the browser, and navigate to it. So the flow was: scan, decode, inject, intercept, navigate. All in maybe half a second.
But the wedge architecture was also the security nightmare. Because the CueCat was essentially a hardware keylogger by design. It could type anything into your computer. And the barcodes determined what got typed. So a malicious barcode printed in a magazine — or just on a website, if someone printed it out — could inject keystrokes that opened a command prompt, downloaded malware, deleted files. Whatever you can do with a keyboard, a CueCat barcode could do.
And the proprietary format meant that nobody could audit what was in those barcodes. You just had to trust that the barcode in your Forbes was only going to open a URL and not type "format C colon" into your DOS prompt.
Walt Mossberg reviewed the CueCat in two thousand for the Wall Street Journal, and he didn't even get to the security issues. He called it "a solution in search of a problem." His core complaint was that the setup was more trouble than just typing a URL. You had to install the software, register with your name and email and household income —
Wait, household income?
Oh yeah. The registration form asked for demographic data. That was part of the business model — they were building a database of who was scanning what. Mossberg pointed out that by the time you found the magazine, grabbed the cat, positioned it over the barcode, and waited for the browser to load, you could have typed the URL six times.
Which gets at the fundamental problem. The CueCat was solving a friction problem — typing URLs is hard — but it introduced more friction than it removed. You had to be at your computer. You had to have the magazine. You had to have the cat. You had to scan accurately. And the payoff was... a webpage you could have reached faster by typing six letters into the address bar.
And the use case got narrower the more you thought about it. If you're reading a magazine on the couch, you're not at your computer. If you're at your computer, you're probably not reading a magazine. The only scenario where it made sense was if you were sitting at your desk, actively reading a magazine or catalog, with the CueCat plugged in and the software running, and you encountered a CueCat barcode and thought "yes, I would like to see the website for this product right now."
That's a narrow window.
It's a pinhole.
By mid-two thousand, Digital Convergence had raised their hundred and eighty-five million. RadioShack had given away something like half a million units. And users were... not scanning things.
The numbers were terrible. The registration rate was low. The scan rate among registered users was lower. Advertisers started asking what they were paying for. And then the dot-com crash hit in two thousand one, and advertising revenue across the entire industry evaporated. Digital Convergence tried to pivot — they pitched the CueCat for enterprise inventory management, for loyalty card programs, for library book tracking. None of it stuck.
The Salon article from July two thousand one is the autopsy. They detailed the security vulnerabilities, the desperate pivot attempts, the layoffs. By that point the company had already shed most of its staff. The servers that resolved CueCat barcodes to URLs were still running, but nobody was scanning.
And then in two thousand two, the company was acquired by a shell company and effectively dissolved. The servers went dark. Every CueCat in the world became a plastic cat that did nothing.
The crash was spectacular, but the story doesn't end with the company's death. The hardware lived on in a second life — one the company never intended.
This is my favorite part. Almost immediately after the servers went down, the open-source community got interested. Because here was a piece of hardware — a perfectly functional optical scanner — that had been given away for free and was now useless. Hackers started reverse-engineering the protocol.
And the first thing they discovered was that the proprietary "encryption" on the barcodes was... not encryption.
It was XOR with a fixed key. That's it. The CueCat would scan the barcode, XOR the data against a hardcoded key to "decrypt" it, and then type the result. The key was extracted from the firmware in about five minutes. Projects like ccdecode and libcuecat popped up, letting anyone decode CueCat barcodes with a regular computer.
So the proprietary lock-in was a single XOR operation. A hundred and eighty-five million dollars of business model, and the security was the cryptographic equivalent of a luggage lock.
But the real hack was on the hardware side. The CueCat's circuit board has a trace labeled J3. If you open the case and cut that trace, you disable the proprietary decoding chip and get raw scan data from the optical sensor. Suddenly the CueCat becomes a general-purpose barcode scanner. It can read standard UPC codes, EAN codes, Code 39, whatever you want — as long as you have software on the computer to interpret the raw signal.
So a device that was deliberately crippled to only read one company's proprietary format could be turned into a universal scanner with a pair of wire cutters and some open-source code.
And people did this. There were tutorials. There were Linux drivers. Someone wrote a plugin that let you scan any book's ISBN barcode and automatically look it up on Amazon. The CueCat had a second life as a dirt-cheap inventory scanner for home libraries and DVD collections.
Which brings us to Daniel's practical question. It's twenty twenty-six. If you find a CueCat in a drawer — and there are still millions of them out there — can you make it work?
Yes, with caveats. The hardware itself is surprisingly durable. The optical sensor is solid-state, no moving parts except maybe the trigger button. Most units still power on and scan. The first hurdle is connectivity. The CueCat uses a PS/2 connector — or sometimes a serial connector — and modern computers don't have those ports. You need a PS/2-to-USB adapter.
And not just any adapter.
Right. Passive adapters don't work. The CueCat draws power from the PS/2 port, and a passive USB adapter doesn't provide the right voltage. You need an active adapter — one with a chip inside that properly converts the PS/2 signaling to USB HID. Those run about ten to fifteen dollars.
So step one, spend fifteen dollars on an adapter for a device that was originally free.
Step two is software. The original CueCat servers are long gone, so even if you scan a genuine CueCat barcode, there's nothing to resolve the URL. But the open-source decoders still work. You can find ccdecode or libcuecat on SourceForge and GitHub. They're old code — some of it hasn't been touched in fifteen years — but it compiles on modern Linux. There are even Windows binaries floating around if you trust random executables from two thousand three.
Which, you know, given that we're talking about a hardware keylogger, maybe don't.
Fair. But if you're running Linux and you're willing to compile from source, you can get the CueCat working as a general barcode scanner. With the J3 trace cut, it'll read standard UPC codes. Without the mod, it'll only read CueCat barcodes — and good luck finding any of those in the wild in twenty twenty-six.
Although someone probably has a PDF of old Wired magazines with CueCat barcodes still embedded.
Almost certainly. And if you print one out and scan it, you'll get the raw URL string. Whether that URL still resolves to anything is a separate question.
The answer to "can you still read a CueCat today" is yes, but you have to want it. You have to buy an adapter, find or compile old software, possibly open the case and cut a trace, and then you have a barcode scanner that's worse in every way than the camera on your phone.
Which is the real punchline. The CueCat's dream — bridging print and digital by scanning a barcode with a device you already have nearby — actually came true. It just took another decade, and the device wasn't a plastic cat. It was the smartphone camera.
QR codes.
QR codes succeeded at everything the CueCat tried to do. They're an open standard — ISO slash IEC eighteen thousand four. No special hardware required. Every smartphone camera can read them natively. The barcodes can encode URLs, contact information, WiFi credentials, payment data. And the business model doesn't require a proprietary lock-in because the value isn't in controlling the barcode — it's in what happens after the scan.
The CueCat's failure wasn't the concept. The concept was right. Bridging print and digital is a real thing that people actually want. The failure was the execution. Proprietary format. Dedicated hardware. A use case that required you to be at your desk holding a magazine. And a registration process that asked for your household income.
It's almost a perfect case study. Every single decision they made was logical from the perspective of capturing value, and collectively those decisions made the product useless. The proprietary format meant no ecosystem. The wedge architecture meant security vulnerabilities. The free hardware meant they needed registration data to justify the cost to advertisers, which added friction. The cat shape meant... actually, I don't know what the cat shape meant.
Branding. They wanted it to be friendly. Approachable. Not a scary piece of technology, but a cute little cat that lives next to your computer.
Instead they got a device that looked like a Happy Meal toy and had the ergonomics of a doorstop. The tail is the handle. Have you ever held something by a tail-shaped handle? It's not comfortable.
I have not held a CueCat.
I have. It's awkward. The center of gravity is wrong. You're gripping this narrow plastic tail and trying to position the nose precisely over a barcode, and the whole thing wants to tilt forward.
The physical design was also working against the use case.
In every way. And yet — and this is what I find fascinating — the CueCat has outlasted almost everything else from that era. The company is dead. The servers are dead. The business model is dead. But the hardware is still out there, still functional, still being hacked and repurposed by people who just like making old things work.
There's something almost admirable about a device that was designed to be locked down and disposable, and instead became an open platform through sheer stubbornness of the user community.
The CueCat was supposed to be a controlled endpoint in a proprietary advertising network. Instead it became a general-purpose tool that anyone could reprogram. That's not a failure of engineering. That's a failure of imagination on the part of the company. They didn't see what they actually had.
They had a cheap, durable optical scanner that could have been sold as a productivity tool. Instead they gave it away as a marketing gimmick and then turned off the servers.
Speaking of people who were there when the servers were still on — Hilbert's been making a face.
Hilbert: The thing nobody talks about is the smell.
The smell.
Hilbert: That translucent blue plastic had a specific chemical off-gassing. Like a new shower curtain left in a hot car. I worked at a RadioShack in the summer of two thousand. We had a cardboard display with a cat-shaped cutout and a stack of CueCats in plastic clamshells. Every time you opened one to put out a demo unit, you got a face full of it. I kept one in a drawer at home for years. Every time I opened that drawer, I got a whiff of the year two thousand.
The plastic outgassing is something I hadn't even considered.
Hilbert: The tail was the handle, and it was surprisingly uncomfortable. You'd think they'd have ergonomically tested a cat shape. We had a demo unit plugged into a Windows ninety-eight machine in the store, and the demo loop was just scanning a barcode that opened the RadioShack website. Customers would walk up and try to use it as a mouse. They'd put it on the desk and move it around. We had to explain, no, you point it at the magazine. They'd say, what magazine?
That's the whole problem in one interaction.
Hilbert: It was. They'd look at the cat, look at the screen, look back at the cat, and you could see them trying to figure out when they'd ever be in a situation where this was useful. Most of them took the free one anyway because it was free. I'd guess ninety percent of them never left the clamshell.
Did the registration forms actually get filled out?
Hilbert: Some did. The managers were told to push it. "It's free, you just have to fill out this form with your email and household income." That was the phrase. We had a script. I probably said it five hundred times that summer. People either laughed or looked at you like you'd asked for their blood type.
Did anyone ever come back and say they loved it?
Hilbert: No.
Did anyone come back and complain?
Hilbert: Also no. Nobody came back about the CueCat at all. It just... left the store and disappeared from the universe. I assume most of them are still in drawers. I know mine is.
Still in the drawer.
Hilbert: Still in the drawer. I don't have a PS/2 port anymore. I'm not going to buy an active adapter for a plastic cat. But I can't throw it out either. It's a perfectly good optical scanner. It just has no reason to exist.
That's almost a eulogy for the whole dot-com era. Perfectly good hardware, no reason to exist.
The CueCat is a punchline, but it's also a prophecy. Every few years someone tries to build a bridge between paper and pixels, and every few years someone makes the same mistake — proprietary hardware, proprietary format, walled garden. The CueCat was just the first and most cat-shaped.
The lesson it teaches is that the best bridge is invisible. You shouldn't need a special device to connect the physical and digital worlds. The device should already be in your pocket. QR codes won because they're just ink on paper, and the scanner is a camera that everyone already carries. No registration form, no driver installation, no plastic tail.
The misconception that sticks with people is that the CueCat was a barcode scanner for regular barcodes. It wasn't. It was a locked-down endpoint in an advertising network that happened to use optical scanning. And that distinction is the whole story.
The other misconception is that the hardware was junk. It wasn't. The optical sensor was fine. The wedge architecture was clever, security issues aside. The plastic shell was durable. The failure was entirely in the business model and the proprietary lock-in. The engineers built a decent scanner. The MBAs wrapped it in a cat and asked for your household income.
Which leaves us with the open question. We're in twenty twenty-six, and we're still building bridges between physical and digital. Every proprietary NFC tag, every augmented reality marker that only works in one app, every smart packaging initiative that requires a special reader — are these just CueCats in waiting? Is the lesson actually learned, or are we just better at hiding the cat shape?
I think the lesson is partially learned. The smartphone camera is the universal bridge now, and most new physical-to-digital systems build on that. But the temptation to lock things down is permanent. Every company that builds a bridge wants to own both sides of it. The CueCat is what happens when you let that impulse run unchecked.
It smells like a shower curtain in a hot car.
Thanks to Hilbert for producing, and for the olfactory detail we'll now never forget.
This has been My Weird Prompts. If you have a CueCat in a drawer and you've actually gotten it working with a modern computer, email the show at show at my weird prompts dot com. We want to see photos.
We'll be back soon.
