Rabbit R1 First Impressions: How I've Been Using the Handheld AI Assistant So Far
The Rabbit R1 can answer questions, play Spotify and record voice memos. But on my first day with it, the camera intrigued me the most.
The Rabbit R1 is a handheld device that you interact with using voice commands.
Jesse Lyu, founder and CEO of Rabbit Inc., compares his company's first product -- the $199 Rabbit R1 -- to a Pokedex. After one day with the device, I'm beginning to understand why.
Just as the fictional Pokedex can identify Pokemon (creatures from the popular cartoon, video game and card game franchise of the same name), the Rabbit R1 can identify items in its environment. Point it at a plant, and it can tell you what kind it is. Aim it at your lunch, and it can tell you what's in it.
Like the Pokedex, it also feels a bit like a novelty so far. The Rabbit R1, despite its tiny size and simple design, claims to do a lot of things. It can call an Uber, order dinner from Doordash, translate conversations, record voice memos, play songs from Spotify and more. Your phone can already do all of those things, but Lyu is promoting the Rabbit R1 as a faster and more natural way to do so.
Rabbit is far from being the only company trying to change the way we interact with devices. There's the Humane AI Pin, another mini-gadget that uses artificial intelligence and a camera to answer questions and help you get things done. That device was criticized by reviewers for its high price, limited functionality and tendency to overheat. The Ray-Ban Meta Smart Glasses also have multimodal AI, meaning the eyewear can "see" what you see and tell you about it.
So far, the Rabbit R1 feels fun, fresh and interesting, but also frustrating at times. It intrigues me, but it also hasn't convinced me yet that there's room for another gadget in my life.
Here's how I used the Rabbit R1 on my first day using it. I'll have more to say in my full review after I've spent more time with it.
Read more: AI Is Changing Our Phones, and It's Just Getting Started
First, what is the Rabbit R1?
The Rabbit R1 is a handheld device about half the size of a phone. It has a 2.8-inch screen, a scroll wheel for navigation, an 8-megapixel camera, 128GB of storage, GPS and accelerometer and gyroscope sensors for motion sensing.
On paper, that makes it sound like a phone from more than 10 years ago, but it's what's on the inside of the R1 that matters. Instead of a traditional operating system with apps, the R1 runs on what the company calls a "large action model." It's software that's been trained to use digital services like a human would, similar to how a large language model provides conversational answers that may sound like a person wrote them. As such, you primarily interact with the R1 by speaking to it instead of swiping and scrolling through apps and menus. However, there is a keyboard for when you need to do things like enter a Wi-Fi password.
The Rabbit R1 is roughly half the size of a smartphone.
Lyu has big ambitions for where this Rabbit OS software can go. In a demo, he's been able to book almost an entire vacation just by speaking a few simple sentences to the R1, such as by asking it to find flights and telling it his preferences. On Day 1, the Rabbit R1 is more limited. Many of the things it can do today feel smartphone-esque, like asking for the weather or playing songs on Spotify.
It takes some getting used to. Years spent tapping, swiping and scrolling will make you forget how to do almost anything else. Sometimes it's refreshing to discover how to use a tech product for the first time, and at other times it's frustrating. Yesterday morning, for example, I asked it to play Taylor Swift on Spotify before leaving the house, just to make sure it worked. The good news: it worked; the bad news: I couldn't get it to stop. It took a few frantic presses of the side button before I finally got it to quiet down.
Visual search is the most interesting feature so far
The Rabbit R1 has a camera, which you can use to ask it questions about your environment.
The Rabbit R1 has a camera, but it's not meant for the typical things you'd use your phone's camera for. It's not about photography, but instead, it's for learning about the world around you. It's pretty accurate for the most part so far. When I pointed it at my salad during lunch, it was able to tell me most of the ingredients.
That's not what I asked. After all, who orders a dish without knowing what's in it? I asked the Rabbit R1 to tell me how many calories were in my lunch. While it couldn't provide the answer I wanted, I was impressed with its response.
It told me it couldn't determine a calorie count because the calories in a salad can vary significantly depending on the ingredients. It would need more details about the ingredients and portion to provide that information, although Rabbit did tell me that it noticed grilled chicken, lettuce, tomatoes and other healthy ingredients. It suggested that I should turn to a nutritional database for a more precise calorie count. Finally, it recommended that I enjoy "healthy meals like this" without "obsessing" over calories.
That may be the best non-answer I've gotten from a virtual assistant. Instead of just saying "Sorry, I can't help with that," as Siri might, it understood my intention, provided the information it could and told me why it couldn't give me the exact answer.
Overall, Rabbit R1's visual analysis worked pretty well for identifying things like plants and characters from pop culture. When describing my colleague's sneakers, the Rabbit R1 got the brand wrong.
This type of functionality isn't new or specific to the R1, and you can already do something like this on your phone through Google's Gemini assistant on Android phones (or the Gemini section of the Google app for the iPhone). It's also very reminiscent of Google Lens, which has been around for years.
How often do you actually use Google Lens to snap a photo and search for something? I'm guessing not a lot. That's what makes the R1 feel different from a phone even if its purpose sounds similar. It's not necessarily about what you're doing with it, but how you're doing it. For better or worse, the R1 forces that type of multimodal interaction by design rather than offering it as an optional input.
The rise of generative AI could make visual search more common, but that surely won't be Rabbit's doing alone. You can't mention search without talking about the 800-pound gorilla in the room: Google. Google is already trying to get us in the habit of moving away from the constraints of words when it comes to search.
Earlier this year, it launched Circle to Search, a feature available on certain Pixel and Galaxy Android phones that lets you search for almost anything on your phone's screen just by drawing a circle around it. No, it's not exactly the same as pointing a camera at something and asking a question about that subject in real-time. It feels close enough to the same idea, and I imagine Google will have even more to say about multimodal search at Google I/O next month.
Other early thoughts
You can press and hold the Rabbit R1's side button and speak into it like a walkie-talkie.
So far, I've used the Rabbit R1 to take voice memos, translate speech from Spanish to English, and answer basic questions about things like weather forecasts. These features work as expected for the most part, and I'll have more to say about them when I've spent more time using them. Your queries and visual searches, along with the services your device is connected to (like Spotify), all live in an online hub called the Rabbit Hole, which you set up when activating the device.
I've also noticed a few hiccups after my first day of use. The biggest of which has to do with ending a task and going back to the Rabbit R1's home screen. Pressing the side button is supposed to put the R1 in standby mode, but there were multiple occasions in which nothing happened -- especially when trying to stop Spotify. This made it frustrating to use at points.
The battery also drained absurdly fast, going from 98% at around 9 a.m. to 34% by 1:41 p.m. It died before I left the office around 5 p.m., even though I had topped it off a bit in the afternoon.
I'll have more thoughts and impressions to share in my review of the Rabbit R1, which I'll be writing after I've had more time with it.
Editors' note: CNET used an AI engine to help create several dozen stories, which are labeled accordingly. The note you're reading is attached to articles that deal substantively with the topic of AI but are created entirely by our expert editors and writers. For more, see our AI policy.