rabbit

first major update of 2026 — DLAM, OpenClaw, and a surprise



2026 is off to a strong start at rabbit. After focusing on improving rabbit r1 over the last year, we've secured new funding, the team is energized, and we're quickly moving forward with big new updates — and a surprise.


Today, we're kicking off this new era by shipping an over-the-air (OTA) update to rabbit r1 that introduces two new agentic experiences, bringing the fast-evolving advances in human-machine interaction straight to your r1.



DLAM - First major update of 2026

DLAM — the first plug-and-play computer controller for everyday users

Our newest AI agent, DLAM, transforms r1 into the first plug-and-play computer controller meant for users of all technical levels—no setup required.


DLAM transforms r1 into the simplest controller for your Windows, Mac, or Linux computer. And the best part is that you don't need to install any software, mess with any complex configuration, or rely on virtual machines. Simply plug your r1 into your computer via USB cable, visit dlam.rabbit.tech, and connect your r1 when prompted. Once you grant screen-sharing access, r1 can see your screen and operate your computer for you. Talk to the agent through the browser or type out your prompts, and DLAM will carry out the actions on your behalf across your operating system, browser, and applications.


We're only scratching the surface, but here are a few examples of what DLAM can already do well based on our testing so far:


  • Simple, everyday actions in the browser: "Go to Amazon and add a 12-pack of Diet Coke to my cart."
  • Multi-app workflows: "Research Tesla's stock performance over the last six months, enter the data in Excel, write an executive-level report in Word, export it as a PDF, and email it to me."
  • Complex software control: "Open Ableton 12, create an audio track and title it 'guitar' and set the bpm to 90."

We even stepped it up a notch and asked DLAM to open Steam and play Civilization VI on its own…and our minds were blown!


We're rolling out DLAM as a free trial today. Keep in mind that DLAM is still very new, and we're discovering new use cases every day. It's already very capable, but like any AI agents out there it will still make mistakes. It's already really good at working through problems on its own, trying different approaches until it finds a solution. It's not fast enough yet, but we'll continue improving it. Try it out and let us know what's working for you.

OpenClaw on r1

you asked for OpenClaw on r1. here it is.

Over the past few weeks, many of you asked the same question: can we get OpenClaw to work on r1? The answer is yes, so we thought, why not?


With today's OTA, OpenClaw access (in alpha) is now available on rabbit r1. The open-source agent formerly known as Clawdbot quickly gained attention for its ability to execute actions directly within users' own environments. What it lacked was native voice interaction on a dedicated hardware device. r1 already fit that role—so we made it happen.


If you've already set up a OpenClaw gateway, you can now issue voice prompts through r1 and have OpenClaw act on them immediately. This integration is early, intentionally flexible, and shaped by how you choose to use it.


Please note:


  • rabbit does not provide support for OpenClaw setup. You will need to set it up on your own first in order to interact with it via r1.
  • All configuration, permissions, and access you give to OpenClaw is done entirely at your own risk and discretion.
  • Since OpenClaw is a third-party project not officially supported by rabbit, we encourage those of you who are already using it and familiar with it to get together and help out folks who might be interested in using this, and help everyone get the most out of it.

And while we're at it, we want to celebrate bringing DLAM and OpenClaw to r1 by offering 15% off a new r1 for the next two weeks, or while inventory lasts. The discount will automatically be applied until 11:59 pm PT on February 12.

now in the pipeline: an entirely new form factor

In addition to continuing to bring more AI capabilities to r1, we're excited to let you know that we're cooking up our next hardware product.


Inspired by the many DIY cyberdeck projects out there, the idea is to create a dedicated "rabbit cyberdeck" for command-line interface (CLI) and native agent use cases. It will be purposefully designed for vibe coders to run things like Claude Code CLI and the upcoming rabbit CLI.


We asked internally what we would like to have as a form factor, and we believe a portable cyberdeck with a really good screen and, more importantly, a hot-swappable mechanical keyboard will help make this device connect with people.


Learning from the industry and our r1 launch experience, we also want this new device to be as adaptable and open as possible. We want you to have the freedom to choose which model or agent to run.


Believe us when we say that this cyberdeck will carry the same rabbit design DNA as the iconic r1, with our signature touch on the CMF.


This time, we want to communicate more transparently with you from the concept phase and gather as much direct feedback as we can from the community. Sign up for updates and join the discussion today in our Discord.


We can't wait to bring this device to life. Happy vibe coding!