The Rabbit R1 is a small, simple mobile device with ambitious plan: to change the way we interact with apps and devices.

Available for pre-order for $199, it’s the first product from a startup called Rabbit, it’s a pocket-sized device with a 2.88 inch touchscreen display, a camera that rotates to face the front or back, and a scroll wheel for navigation, as well as far-field mics and a push-to-talk button with support for voice input. What does the Rabbit R1 actually do though? A lot… and not much, depending on how you look at it.

So here’s the thing about the Rabbit R1: it’s smaller than a phone, has a SIM-card slot and USB-C port, and… isn’t really designed to be a phone.

Instead of Android, iOS, or even a mobile GNU/Linux distro, it runs a new operating system called rabbit OS that places an emphasis on a type of AI that the company calls “Large Action Models,” or LAM.

In some ways, you can think of LAM as working the way “skills” do on voice assistances like Amazon Alexa. You don’t need to navigate to an app store and then download and install an app before you can start streaming music. You just ask Alexa to do it for you.

Similarly, Rabbit says it’s trained its services to run apps for you on the company’s “secured cloud,” so that you can do things like stream music from Spotify, request a ride from Uber or Lyft, order food from GrubHub or DoorDash, and so on.

This works because Rabbit basically trained its AI on how users interact with those apps, allowing everything to happen in the cloud when you just ask it do perform an action.

Want to do something that Rabbit OS doesn’t already know how to do? You can log into a web portal and fire up a virtual machine to train the software to perform new actions. The portal is also where you’ll login to Spotify, Uber, DoorDash, etc to make sure your requests are linked to your own account. For privacy purposes, the web portal will redirect you to the login pages for each of those apps – it doesn’t store your passwords on its own servers.

I was pretty skeptical of the whole concept until I found out that the web portal is called Rabbit Hole, and now I have to say I’m kind of rooting for the company based on that fact alone.

But honestly, the Rabbit R1 and Rabbit OS do seem to be trying to solve a problem that may or may not really exist. It’s a new way to interact with existing apps and services that you may already be using as well as a virtual assistant (accessible by asking questions with your voice or by shaking the R1 to reveal an on-screen virtual keyboard).

And while a simple, primarily voice-controlled user interface is nice, I’m not sure how much demand there is for a stripped down way to interact with existing apps.

That said, it could have some appeal for less tech-savvy folks who don’t already have smartphones or tablets and/or find it challenging to navigate through an endless sea of apps to get things done. And maybe one day Rabbit could enable new forms of experiences that aren’t possible today.

Since most of the work is done in the cloud, the Rabbit R1 doesn’t need a lot of processing power inside. According to The Verge, it has an unspecified 2.3 GHz MediaTek processor, 4GB of RAM, and 128GB of storage. And the $199 price tag isn’t bad for a mobile device that was designed with the help of Teenage Engineering (which isn’t exactly known for wallet-friendly products).

But I am a little dubious of Rabbit’s claim that there’s no subscription required to use the device. Because unless Rabbit plans to start throwing ads at users, it’s hard to see how they can grow their user base and keep their cloud services online in the long term by selling relatively inexpensive mobile devices.

Rabbit’s R1 is up for pre-order now, and should begin shipping in March, 2024.

Support Liliputing

Liliputing's primary sources of revenue are advertising and affiliate links (if you click the "Shop" button at the top of the page and buy something on Amazon, for example, we'll get a small commission).

But there are several ways you can support the site directly even if you're using an ad blocker* and hate online shopping.

Contribute to our Patreon campaign


Contribute via PayPal

* If you are using an ad blocker like uBlock Origin and seeing a pop-up message at the bottom of the screen, we have a guide that may help you disable it.

Subscribe to Liliputing via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 9,536 other subscribers

Join the Conversation


Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  1. I dream about normal linux on similar old sony UX palmtop (palm os)
    small similar camera frond and back, good keyboard, pen stylus long working time network

  2. there is nothing new in it, just some device with google like assistant more like smart speaker or something like that, but more mobile.

  3. I read the article and still can’t understand anything exept the “solutuon looking for a problem” part.

    1. ChatGPT* says:
      Basically, it’s a cell phone, but instead of running any software you recognize, you’ve got some mystery meat neural network application that takes in voice commands and finger waggles and outputs macros. These macros control a virtual machine that’s running android or maybe other stuff.

      *not really.

  4. A “mobile companion”? Well, I guess $199 is cheaper than a marriage, so maybe it’s a win for some basement dwellers.

  5. As much as I love weird and quirky little devices (and rabbits!), this feels like a solution in search of a problem. Hard pass, especially at $200.

    1. May work for someone with mild Alzheimers, but they may forget where they put the device. That could be helped by having device sown into clothing.

    2. You know, it occurred to me that maybe the biggest, or at least, most active group of people using this thing aren’t actually going to be people at all.
      Since it’s operated by voice commands, people can hook a TTS engine up to a LLM chatbot and have the chatbot issue instructions to the R1. Or, you know, like anything running an Alexa interface. Except there’s (theoretically) even more things this can do than Alexa or Siri, thus more things that can be automated…if you can get enough vram.
      Get ready for biometrics to become mandatory for interacting with a lot more interactive computer services and government imposed limits on how much vram you can get in a graphics card!

  6. Brad said: “But I am a little dubious of Rabbit’s claim that there’s no subscription required to use the device. Because unless Rabbit plans to start throwing ads at users, it’s hard to see how they can grow their user base and keep their cloud services online in the long term by selling relatively inexpensive mobile devices.”

    I agree with you Brad. They keep pushing the “no subscription” model and “privacy” but never clearly state how they are going to monetize what looks to me as a service with heavy-lift on the back end. So unless every single thing you buy through the “Rabbit Hole” has a hefty surcharge, and/or they are being paid to spy on you 24×7, I don’t see how they get paid.

    $199 – no subscription required

    $199, no subscription.

    25 minute video pitch – it only mentions the one-time $199 device purchase price:

  7. @liliputing_

    This is triggering my BS detector. They're taking $200 now to ship around Easter, but their FAQ says sales are final and non-refundable 14 days after purchase.

    Seeing Teenage Engineering attached is a reassurance, but hardware is hard. Look how long it took for Panic to ship the Playdate, with the same price and design partner, and a much better defined product.

    I don't think it's a total scam, but I'm wary. I'm not a fan of AI generally, but there is something compelling out of a tool that could opt out of dark patterns and bad UI.

  8. If they let you download the server and just had you pay for a reverse proxy service I’d trust it a lot more. But even then I still wouldn’t use it because yelling commands to nobody in public is going to get you a lot of unwanted attention. As it is, this is basically the logical extreme of how Big Data approaches services. All the stuff that doesn’t require a Nvidia Tesla card (because that costs a lot BUT NOTHING ELSE WORKS) happens on the server, and all the heavy neural processing that would happens on the device instead, and both ends are proprietary, and presumed spyware.
    Not that, you know, its socially acceptable to care about things like that, you monster, WE SEE YOUR GUILT.
    In any case, they’d sell more units if they could cram it into a device the size of its screen.

    1. I think their goal is to make it seem like everyone has Tourette’s.