A stacked box that turns a pile of clothes into a pre-inventoried parcel.
The concept is a two-chamber box a household can 3D print, borrow from a neighbourhood store, or receive as a plain cardboard unit. You put garments into the lower chamber and it folds them. The upper chamber measures dimensions and assesses the garment, fabric and condition, without cutting or tearing anything. What you take out is a folded, inventoried parcel you set on your doorstep.
The reason to want a physical device rather than an app is friction. Photographing, measuring, describing and listing a garment takes fifteen minutes of attention per item, which is why the garments in the cupboard stay in the cupboard. A box you drop things into asks for close to none.
The reason to want it at all is that nothing else produces this data. Every consumer resale platform checked, ThredUp, Vinted, Depop, Vestiaire Collective and eBay, still has the seller type measurements in by hand, and none documents automated condition grading. The measuring chamber is not a convenience on top of an existing capability. It is the capability.
A flat garment photographed against a known reference plane gives chest, length, shoulder and sleeve measurements to within the tolerance second-hand listings actually use. This is ordinary computer vision on a fixed rig, and the fixed rig is the easy case: lighting, distance and background are all controlled by the box.
Near-infrared spectroscopy distinguishes the common textile fibres non-destructively and is deployed at industrial scale in sorting facilities today. Feasibility is settled; price is not. The cheapest confirmed route is the Matoha handheld at GBP 1,149 of hardware plus GBP 199 a month for API access, which is a small-business tool rather than a household one. trinamiX, Picvisa, Fibersort and the Refashion sorting line are all enterprise, quote-only, and physically installed in a facility. So this capability is known to work and, at consumer prices, currently unavailable, which is exactly the case for a shared neighbourhood unit rather than one box per household.
Not established
Pilling, thinning and colour loss are visible to a trained eye and to a good camera. What is missing is the shared scale to report them on, which is the same gap the fabric score has. The hardware could measure something today; there is no agreed something to measure it against.
Not established
This is the unsolved part, and it is unsolved in the strong sense. Deformable-object manipulation is a live robotics research problem: a garment has effectively unlimited configurations, self-occludes, and has to be grasped and re-grasped before its shape is even known. Two well-funded consumer attempts have failed commercially. Presenting a folding box as a near-term build would be dishonest, so it is named here as the open engineering gap that decides whether the physical device is a product or a thought experiment.
Both commercial attempts at consumer laundry folding failed, and neither failed for want of money or attention. Laundroid ran for about a decade inside a company that ultimately went bankrupt. FoldiMate demonstrated repeatedly at CES and never shipped. The machines that did work were room-sized, slow, and unreliable on anything that was not a rectangular shirt.
The underlying difficulty is that a garment is a deformable object with no fixed shape. A robot has to first discover the configuration it is in, which usually means picking it up and letting it hang, then re-grasp it at points it can only infer. Rigid-object manipulation is largely solved; this is not that problem.
The academic work has kept moving and has not shipped. SpeedFolding, GarmentPile, DexGarmentLab and Instant-Fold are all real published results between 2022 and 2026, and the most recent of them is still an arXiv paper. Five-plus years after the second commercial failure, that is the state of the art.
What follows from this, and it is a design conclusion rather than a disappointment: the measuring chamber should be shippable on its own. It delivers most of the value, since the inventory record is what unlocks the listing, and it does not depend on the hard half. A device that measures and does not fold is a real product. A device that promises folding is a research programme.
Open geometry, printed in parts, assembled around a phone or a cheap camera module. Lowest unit cost, highest assembly effort, and the version that makes the design genuinely forkable.
One good unit shared across a street or a building. Best hardware per euro, and it fits the regional collective pattern that the second-hand work already leans on.
A folded reference plane with printed markers, and the household phone does the imaging. Almost free, ships flat, and covers the measuring case completely.