A weekly outfit planner for households that tracks which clothes are clean, in the wash,
or needed next – so laundry happens before anyone runs out.
Role: Solo – research, product, UX/UI, build
Timeline: 1-week sprint
Tools: Figma, Lovable
Deliverable: Working web app
I set out to solve an everyday household problem end to end, from defining it to shipping a working product in one week, designed in Figma and built with Lovable.
The core problem isn’t laundry. It’s not knowing what’s clean, what’s in the wash, and whether there’s enough for the days ahead. T-Short starts from a weekly outfit plan. Once the household decides what everyone will wear, the app tracks each item’s status and shows what needs washing, and when – before anyone runs out.
Every week, people make small decisions with information they don’t have:
Which clothes are clean and ready to wear?
What’s still in the laundry cycle?
Do I have enough for the next few days?
When should I start the next load?
Each question is simple. Together, they create constant mental load, because the answers keep changing and nothing keeps track of them. So people find out about shortages only when they reach for something that isn’t there.
Existing habits manage the laundry process. None of them answer the question users actually care about:
What’s available right now, and what will be available when I need it?
That shifted the product from managing laundry cycles to managing clothing availability.
With one week and a working product as the goal, I kept research lean. I started from my own experience as a user, held informal conversations with potential users, and analyzed online forum discussions about managing household laundry. Three patterns came up – all of them reactive.
T-Short treats each clothing item as a unit with a status that updates as it moves through its lifecycle: in the wardrobe, worn, in the wash, back in use. This gives households a live picture of what they have now, and what they’ll have soon.
Three capabilities work together:
The core of the product. At the start of the week, the household plans what each member will wear each day, with optional AI suggestions. Every morning, the answer is already there.
What's clean and what's in the wash, per household member, updated automatically from the plan.
What needs washing, and when, so the planned outfits are clean in time.
The system only works if item statuses stay accurate. Reviews of existing wardrobe apps pointed to the same barrier: users give up during setup, because adding clothes one by one is exhausting. So I treated onboarding as a core feature, not a one-time form:
For generic items like socks or undershirts, one photo plus a quantity replaces dozens of entries.
The trade-off: this flow works best when users already have photos of their clothes.
After setup, the weekly plan keeps the data accurate. The day after a planned item is worn, its status automatically changes to “dirty”. The plan is the input – users never update items one by one.
I designed around the full lifecycle of an item rather than individual screens. Three decisions shaped the product:
Users think “where is my blue shirt?”, not “which screen am I on?”. So every flow starts from the item and its status.
No two households do laundry the same way. Instead of enforcing a process, the product follows how families already organize clothes – by household member and by category.
Users shouldn’t have to remember what’s in the wash. Availability appears exactly where a choice is made, like when planning an outfit.
I built the product in Lovable through short iterations: define the goal, write the prompt, review the output, refine.
I designed the first screen in Figma and used it as the reference for everything Lovable built. I reviewed every new screen against it, down to shadows, strokes, corner radius and icon style. Each fix was set as a rule across the whole product, not just the screen in front of me.
The navigation bar was the clearest test. Lovable’s first version came out solid black, which clashed with the feeling I was designing for: clean, airy, light as soap bubbles. Asking for a different color wasn’t enough – the component still felt off. So I sketched the bar in Figma, chose matching icons, and gave Lovable exact specs. This time, the result matched the design.
The lesson: whenever an instruction left room for interpretation, the output fell short. Lovable handled execution. The design, the thinking and the method stayed mine.
Every household had its own laundry logic – by person, color, fabric or urgency. Forcing one workflow would have meant asking users to change their mental model before they saw any value. Organizing by family member and category let the product fit in first.
A system that tracks state is only as good as its data. Asking users to update every item was never going to last. Planning the week is something families want to do anyway – so the plan became the data entry.
A fix only held when I set it as a rule for the whole product, not a patch for one screen. The same thinking that keeps a design system consistent kept the AI output consistent.
Next, I’d give households more control when reality doesn’t match the plan: changing an item’s status manually, and customizing the status rules in settings – for example, items that can be worn more than once before washing. I’m also planning usability tests with households to see how the weekly plan holds up in daily life.
Longer term, the real answer is removing manual input altogether – for example, a washing machine that recognizes what goes into it.