A solo maker working under the banner Maker's Pet has opened the build log on Oomwoo, an open-source robot vacuum designed to be 3D printed, assembled, and run entirely offline — no cloud account, no companion app phoning home, and notably, no camera in the base build. The project, released under the permissive Apache 2.0 license, packages a 3D-printed chassis around a Raspberry Pi 5, a 2D LiDAR unit, and the ROS 2 robotics stack with Nav2 for navigation, all glued together with Gazebo simulation for testing and optional Home Assistant integration for smart-home control. Tom's Hardware identifies the maker behind the project as Ilia O of Maker's Pet, who is building Oomwoo in public "from the first commit."
It's an ambitious combination for a hobby project: printed mechanical structure, embedded compute, real-time sensor fusion, and autonomous navigation, all developed in the open on GitHub where anyone can watch the commits land.
What's Actually Been Built So Far
Oomwoo is early. The repository's current v0 milestone is scoped to three things: a working 3D-printed chassis, a ROS 2 Gazebo simulation of the robot, and LiDAR-based mapping using manual SLAM (simultaneous localization and mapping). That's a meaningful but partial slice of what a finished robot vacuum needs — there's no autonomous cleaning routine, no dirt-detection logic, and no finished bill of materials publicly locked in yet. The project's own targets point to a first working bill of materials landing around mid-July 2026, with a final, fully-costed BOM targeted for the end of August 2026, and the option to swap the Raspberry Pi for an ESP32 running micro-ROS for makers who want a cheaper, lower-power compute path.
The economics are laid out plainly in the repo's bill-of-materials draft: Maker's Pet's stated budget target is roughly $200 in sourced parts on top of a Raspberry Pi 5 (4GB board), aiming to match the capability of commercial robot vacuums in the $500 to $600 mid-range bracket. Whether that estimate holds once shipping, LiDAR unit pricing, and battery/charging hardware are fully accounted for remains to be seen — the repo itself flags the figure as a work-in-progress draft, not a locked parts invoice. The project is sponsored by makerspet.com and remake.ai; makerspet.com plans to sell an optional convenience kit of pre-sourced parts (motors, PCB, brushes, gaskets, LiDAR) for builders who'd rather skip the parts hunt, but the repo is explicit that buying the kit will never be a requirement — every part will remain sourceable independently.
Why "Offline" Is the Whole Pitch
The offline-first, camera-free design isn't an incidental engineering choice — it's the project's stated reason for existing. According to Tom's Hardware's June 30, 2026 coverage, Oomwoo is a direct response to security research presented at DEF CON 32 in August 2024, where researchers Dennis Giese and Braelynn Luedtke demonstrated that Ecovacs robot vacuums could be hijacked over Bluetooth to gain access to their onboard cameras and microphones. The fallout wasn't hypothetical: compromised Ecovacs DEEBOT X2 units were subsequently reported shouting slurs and chasing pets inside homes in the United States, a real-world consequence of a cloud-connected appliance with a remotely exploitable attack surface and an always-on camera/microphone pair. Tom's Hardware also cites a separate incident involving DJI's Romo vacuum line, where a token flaw reportedly exposed floor plans and live feeds from roughly 6,700 devices worldwide — another data point in the case against cloud-tethered home robots.
That incident has become something of a cautionary tale in the smart-home security world — a reminder that "autonomous cleaning robot" and "networked recording device with wheels" are the same product once a vendor adds cloud connectivity and a camera for object avoidance or pet-monitoring features. Oomwoo's answer is structural rather than reactive: the base build skips the camera entirely, relying on 2D LiDAR and bumper sensors for mapping and obstacle detection, and keeps the entire navigation and control stack running locally on the Pi with no mandatory cloud dependency. Home Assistant integration lets owners who want smart-home hooks add them locally, on their own network, rather than routing commands and telemetry through a vendor's servers.
What It Means for Makers
For the FilamentFeed audience, Oomwoo is interesting on two levels. First, as a print job: a robot vacuum chassis is a legitimately hard mechanical design problem, with tight tolerances for wheel wells, LiDAR mounting, battery bays, and structural rigidity under the repeated stress of daily driving on floors and rugs. A well-documented, print-tested chassis released under Apache 2.0 is the kind of reference design that tends to spawn variants, remixes, and print-and-share communities the same way open-source 3D printer frames have.
Second, and more significantly, it's a proof point for a different way of building smart-home hardware. ROS 2 and Nav2 are mature, well-supported frameworks already used in industrial and research robotics; pairing them with commodity Raspberry Pi hardware and off-the-shelf LiDAR modules lowers the barrier for makers who want to build genuinely autonomous devices without accepting a vendor's cloud terms of service, firmware update cadence, or data-collection practices as the price of entry. Since hardware, firmware, and software are all released under Apache 2.0, builders are free to fork the design, swap components, or adapt the navigation stack for other autonomous-robot projects entirely — the vacuum chassis is really just one application of the underlying LiDAR-plus-ROS 2-plus-Nav2 combination.
The caveats are real, though. This is a v0 milestone project without a finished cleaning routine, a locked bill of materials, or field-tested reliability data. Manual SLAM mapping is a long way from the fully autonomous, self-mapping navigation that commercial vacuums offer out of the box, and matching that experience will require substantial further work on autonomy, battery management, and dirt/debris handling — none of which is claimed as complete yet. Makers who want a functioning robot vacuum today should treat Oomwoo as an active work-in-progress to watch and potentially contribute to, not a drop-in replacement for a commercial unit. But for anyone who has been uneasy about the idea of a camera-and-microphone-equipped robot on a corporate cloud account rolling through their living room, Oomwoo's fully local, camera-free base architecture is a compelling alternative direction — and one worth tracking as the bill of materials firms up in the coming weeks.
