Pane OS is live and accessible at the link below. Open it in a browser, press Summon to begin, and swipe right on the pane at any time to return home.
Pane OS is a speculative home OS concept for the 2040s. It takes the form of a fluid glass panel that lives on the wall, moves between rooms, and splits itself to serve multiple family members simultaneously. The brief asked for a smart home experience set fifty years in the future. The response was not a collection of connected devices but a single surface that makes the home feel genuinely alive.
The prototype is a fully interactive HTML experience deployed via Vercel. It runs in any browser, requires no installation, and works on desktop and mobile. It includes eleven screens spanning entertainment, communication, home management, energy monitoring, creativity and play. Four real photographic room environments show the panel in different domestic contexts. Voice search, real-time clock, time-aware suggested content and camera access via FaceTime all function as live interactions rather than simulated ones.
The prototype makes the case for Pane OS across four arguments.
It is for the whole family. The video narrative moves deliberately from parents to friends to children to grandparents. Each screen visited belongs to a different person in the home. The child lock in the bedroom, the drawing canvas, the Minecraft screen, the energy dashboard for the adults and the FaceTime call for the grandparent all exist to show that this is not a product with a single target user.
It is genuinely speculative. The energy screen showing live solar generation and carbon offset, the split mechanic sending a second panel to another room, the ambient intelligence of the time-aware home screen all sit comfortably in 2043 without feeling disconnected from the present. The technology is extrapolated rather than invented, which makes the concept believable rather than fantastical.
It is considered as a design object. The glass language, the typographic restraint, the sound design, the spring physics on every animation and the Japandi photographic backgrounds were all deliberate choices made in response to research and refined through iteration. The aesthetic coherence of the prototype was consistently noted as a strength in testing and critique.
It is built rather than drawn. Every interaction in this prototype is real. The pane springs, the voice listens, the camera activates, the sliders drag, the gallery opens and the drawing animates. None of this is possible in a Figma prototype. The decision to build in code was a considered one and it produced something that communicates the concept more honestly than any static mockup could have.
Looking back at where this project started, the distance travelled is significant. The early research pointed toward a concept that lives between furniture and technology, between architecture and software. The moodboard referenced Flubber movement, the GLaDOS idea of a building that thinks, the light-through-glass quality of something both present and transparent. Those references shaped every visual decision that followed, from the 0.5px border highlight that makes the glass readable to the spring overshoot on the summon animation. The outcome was not accidental. It was the result of research being taken seriously.
The decision to prototype with Claude rather than Figma deserves a moment of reflection here because it was the most unconventional choice in the project and arguably the most important one. Design teams are increasingly working this way, using AI as a development collaborator to produce interactions that prototyping tools cannot replicate. The experience of building Pane OS through that process taught me things about interaction design that Figma never could. When you have to describe exactly how something moves, exactly what it sounds like, exactly what happens at the moment of release, you are forced to have opinions about things that a drag-and-drop prototype lets you leave vague. That specificity made the concept better.
If the project continued, the priority would be real room footage with the prototype composited in, which was the original intention before time and tool constraints made it impractical. The photographic backgrounds in the current prototype are a significant improvement on the CSS gradients they replaced, but nothing would communicate the concept as powerfully as a real wall with a real pane of glass appearing on it. That remains the version of this project worth making.