Rabbit OS3: Cross-Platform AI Agent Moves to Software

What happened

Rabbit, the company that tried to bypass mobile apps with dedicated AI hardware, is launching OS3, a cross-platform AI agent that runs on the screens people already use. Wired reported the launch and framed it as a return to the company's core idea, this time delivered as software rather than a separate device.

The move is a notable change of direction. Two years after its first hardware push, Rabbit is betting that an agent embedded in familiar devices can do what a standalone gadget struggled to do: fit into existing habits without asking people to adopt something new.

Why it matters

The shift points to where AI assistants may be heading. Instead of being a destination users must visit, an agent that operates across platforms becomes a layer sitting above the apps, websites and services people already rely on. For organisations, that changes the question from "Do we have an app?" to "Can our service be understood, navigated and completed by an agent acting for the customer?"

It also raises the stakes on trust. An agent that moves across screens has to be clear about what it can see, what it can do and when it is acting on someone's behalf. Those are design and governance decisions as much as technical ones, and they will shape whether people feel in control or quietly unsettled.

The Renascence take

Most coverage of this launch will focus on whether Rabbit has finally found a product that works. The more useful reading is that the company has conceded a point many service teams have been slow to accept: people do not want more interfaces, they want outcomes with less friction. The behavioural principle underneath is simple, since the path of least resistance wins, and an agent that lives on existing screens removes the cost of switching.

Rabbit's pivot is less a comeback than a warning that the device was never the point. If an agent can act across the screens customers already use, the experience they judge is the one the agent creates, not the brand's own app. Operators should audit their journeys now, asking where an agent could complete a task on a customer's behalf and where it would fail, confuse or overstep. The organisations that win will be those that design for consent, transparency and recovery from error, not just for speed.