Most companion robots on the market work like this: You speak to them, the audio goes to a server somewhere… California, maybe Shenzhen, maybe…you get the point. That audio then gets processed and comes back as a response. The conversation is logged somewhere. It’s analysed. It may be used to train future models. Somewhere, a system knows what you said, when you said it, and to whom.
This is not fearmongering and not rage baiting, either. That’s simply the standard business model, and it is openly stated in the terms of service and privacy policy.
For a smart speaker or a voice assistant, most of us have decided that the trade-off is acceptable. Yes, I also use certain assistants, and you bet I have subscriptions to pretty much every GPT company out there. And I even use the fingerprint sensor on my Mac!
But then, a speaker is a tool. A companion robot, on the other hand, is something different. And anyone who says robots in private households are just “some new tech” is lying. Your companion robot is the one your child will tell their worries to at bedtime when they don’t want you to hear what they say. It’s the device your elderly parent talks to when no one else is home.
We quite literally design them, explicitly, so you become attached to them and get closer over time.
That context changes everything.
We know we build something that people genuinely trust, but unlike others, we will never be casual about what we do with that data.
Why the privacy-focus of Reegy’s robots?
We strive for privacy focus in our products because I – and Reegy as a whole -strive for one thing above all else: Integrity.
To me, privacy is actually just part of a “grander” idea: Building products and companies that are beneficial for society, that are built without exploiting anyone or anything. While this might sound a bit like a utopia, it has worked out great so far.
One day, we will all have robots as companions, in our house, maybe in our bathroom, during family night, or next to the drawer in our bedroom. We want to be at the forefront when that happens, but we absolutely have no desire to be part of your life through our robots.
Why Being European Matters Here
I’m a mechanical engineer with an engineering degree from Germany. Reegy is a German company. Our companion robots are being designed and built in Europe. We love that part and even though Silicon Valley sounds great for people like us…it’s not for us.
Europe has GDPR. The legal culture here treats personal data as a priority that belongs to individuals, not corporations. It has regulatory pressure on surveillance technology that doesn’t exist in the same way elsewhere, especially not at other startup hotspots.
Your Reegy robot is an EU-native.
In my opinion, a company that builds a companion robot and monetises the emotional data it generates is doing something wrong, even if it might be technically legal. And I think personal data flowing out of Europe to other places in the world is a problem. We want to be different.
Reegy is different, and my answer to that problem.
What “Privacy-First” Actually Means at Reegy
Before the first line of code of Reegy was ever written, and before we put the first microcontroller in a casing, I had already spent days trying to figure out how to build a robot that is 100% personal and private to whoever owns it. Not just in a “we will handle your data carefully” way or out of compliance reasons, but rather: We have a robot and whereas others want to use it for training on data or connecting to as many services as possible, we will do the exact opposite. We make our robots feel like a person, or a friend. And you wouldn’t want your friend to tell the whole world in real-time what you just did or said.
Here’s what that means in practice:
AI processing happens on-device by default.
The “brain” of our robots, the primary intelligence core (PIC), runs completely locally on the robot itself. It is a separate, really, really cool and sophisticated piece of tech that stores your robot’s memories of you. Your voice, your conversations, your memories: none of that leaves the device unless you explicitly choose otherwise. There’s no opt-out you have to find buried in a settings menu, and no “we’re training on your private conversations, that’s okay, right? See ya, bye”-toggle. Local-first is the default, not the fallback.
Nohmu’s memory is yours.
The persistent memory that makes your Reegy robot a companion, so the memories of your shared history, your preferences, and the things it’s learned about you are stored locally and encrypted. We can’t access it. We don’t want to. If you delete it, it’s gone. No need to call us for help either, we literally don’t know what’s on there and cannot get it back. (We have super good support in all other areas, though!)
Cloud AI is strictly opt-in.
We know that cloud-based models like Claude or GPT are genuinely more capable for certain tasks than what we can offer right now. We’re not pretending otherwise. So we’re building the option in. If you absolutely want, you can choose to connect to a GPT of your choice so your robot has a second (actually a third, to be precise) brain unit for some really super-power-like abilities. But it will always require a deliberate, informed choice by the user. The moment you enable it, we tell you clearly what that means. The moment you disable it, it stops (but remember what I said at the beginning, whatever you shared before is already on their servers).
Which brings me to the next point…
Reegy SuperbrAIn is also optional
Reegy’s cloud AI – lovingly called the Reegy SuperbrAIn – is optional and not required to use your robot. You can use your unit, including all the extra modules and modifications you want, without ever touching it. However, here is what makes Reegy SuperbrAIn really really good in terms of privacy:
- We host it in Germany, a country with some of the strictest data privacy laws in the world. Not only does it fall under the EU’s really strict privacy laws, but it also supersedes those with some extra strict ones. It’s a pain in the butt to implement but we are all here for it.
- It runs on our own servers. Our business is building and selling you the best robots for home, family, and business out there. Therefore, we have absolutely no reason to hide how it’s built or trained. If you want to know something, just ask us!
- It’s really, really powerful. Yeah, our SuperbrAIn isn’t called that for no reason. Our models are state-of-the-art and let your robot expand its conversational and world skills exponentially.
What This Means for You
If you leave your Reegy robots at home with your kids, they aren’t data providers or part of a training dataset. If you talk to them about something personal, that conversation stays between you and the robot. If you decide tomorrow that you don’t want it anymore, the memories it holds are yours to keep or delete (remember though what I mentioned about not being able to get them back).
That’s the kind of robot I wanted to build. That’s the kind of company Reegy is going to be.
We’re still early. Our robots are still in active development, and there’s a lot of work ahead before it’s ready to come home with you. But the architecture is set. The principles are locked. This is not a promise we’ll make until investors or “growth” demands we break it.
If that sounds like a robot you’d trust in your home, we’d love to have you follow along.
