Maket debuts Auto-Complete for generating residential floor plans

Maket has released Auto-Complete, a feature that turns a partial floor plan into a finished residential layout while leaving the rooms a user has already placed exactly where they are. The starting point can be minimal: a drawn outer shape, a few walls, or a single bedroom positioned roughly where it should sit. From that input, Maket fills in the remaining rooms and returns a complete, dimensioned plan in minutes, rather than asking for a full description before anything appears on the canvas.

The method targets a familiar gap in generative design. Tools that build a plan from one prompt tend to treat every layout as all or nothing, which means the parts a person already feels sure about can get reshuffled on each new pass. Auto-Complete reverses that order. The fixed rooms stay fixed, and the system designs around them, so early work starts from what is already decided instead of resetting every time.

Once a layout lands, the same workspace carries it forward. A generated plan can be opened in 3D to move through the space, and renderings can be produced from a short text prompt alongside optional reference images, taking a draft from a flat layout toward a realistic look without a separate tool. Generation, refinement, and visual direction stay on one canvas, which is the structure behind Maket's v2 platform.

Auto-Complete is available inside the Maket web app, aimed at homeowners, builders, and real estate professionals who need to move quickly through the early design phase. Maket starts on a free plan with no credit card required, with some advanced capabilities reserved for paid tiers.

Maket presents itself as an AI floor plan studio built to make residential architecture reachable for people without CAD software or formal design training, including help to navigate zoning codes while planning. More than one million people used the first version of the platform, and the current v2 release brings plan generation, iteration, and visualization onto a single canvas. Auto-Complete fits directly into that direction, lowering the starting effort so a partial idea is enough for the system to propose a complete plan.