The TuckBuddy app icon: a teal square with a light diamond and a dot at its center.

I built and shipped an iOS home inventory app in under four months

TuckBuddy tracks what we own across two homes. I designed it, ran it like a product team with AI bot-workers in the roles I didn’t have, and shipped it free.

Role
Founder, designer and product lead
Team
Me and seven AI bot-workers
Platform
iPhone, iOS 17.6 and later
TuckBuddy 1.0 Home for Maple Street: a Last tucked card showing a Dutch oven, then the home's spaces.
01Overview

The problem

Most home inventory apps either bait and switch, requiring a purchase after a small taste of the product, or overspecialize for clothes collectors, insurance manifests and the like.

What I did

I researched competitors, wrote the requirements and roadmap, designed every screen, and directed a team of AI bot-workers I designed to build it with me.

The result

TuckBuddy 1.0 is on the App Store, free with no item cap. It keeps your data private in your own iCloud, is accessible by default, and meets collectors and generalists where they are.

02The problem

I wanted to inventory my life, and the apps I found punished people for trying

The pattern across competitor reviews was the same everywhere: people spent an evening cataloging, hit an item cap, and felt tricked. I wanted the opposite promise, so TuckBuddy is free with no cap.

Competitive analysis: a positioning map with Tuck alone in the corner for one focused job for the people at home, then head-to-head cards against Sortly, Encircle and HomeZada.
Competitive analysis. Drafted with Claude, edited and decided by me.
Naming study: three plain-word name candidates, Cubby, Shelf and Tuck, each with an icon and notes.
Naming study. Drafted with Claude, edited and decided by me.
Brand voice study: phone screens showing how the word tuck carries through onboarding, the add flow and empty states.
Brand voice study. Drafted with Claude, edited and decided by me.

Real people shaped what the product needed. A couple living across two homes needed to know which home a thing is in so they stop buying duplicates (That’s me. I live in SF and housing costs make this a surprisingly common scenario). Partners or housemates managing a home inventory together. Someone who swaps out seasonal decor with multiple bins in multiple locations. At the far end, a minimalist wanted to type “shirt” and be done, so every detail had to be optional.

I wrote a laddered roadmap where each phase had to be worth shipping even if I never got to the next. Phase one was a private, single-user app: homes that hold spaces, spaces that nest as deep as you like, items that go anywhere, and search that tells you where a thing is. Sync came next, and sharing after that.

03Key decisions

Three calls shaped the app people use today

Hallway testing broke my category system, so I rebuilt it around tags

I started with categories and subcategories, with open vocabulary tagging on top. Clothing > Shoes would get most people there. Testing ended that fast. Shoes alone needed house shoes, slippers, flip-flops and sandals, and people said picking from a fixed list took so long they couldn’t picture using the app.

I cut the categories from 22 to 11, each deciding which details appear. Subcategories became suggested tags you can stack, so one item can be formal, workwear and wool at once. Tags you use stay up front and one-offs fade until you need them again.

The details and tags editors on a pair of boots

The app is gender neutral, and respects the name you use

Gender expansive people are everywhere and in my life, and they shouldn’t have to make unneeded choices just to do chores. Size is just a field. We ask for the name you use because when sharing and possibly payments come along, we aren’t going to use the legal name on your government documents just to tell you who moved the umbrella to the hallway closet.

Adding an item needed a small reward, so I designed The Tuck

Inventory can feel like a chore. I first wanted each new item to fly to the exact place it lives, but you can add anything to anywhere, so that place is often off screen. Instead the item’s icon or photo hops into the confirmation at the top of the screen, which also tells you where it landed. Also, I didn’t want to insert mechanics that might enable unhealthy behaviors like hoarding, so no streaks or thirsty guilty prompts anywhere in the app.

The Tuck
04What fell short

Two things went wrong and I will run my next AI build differently

I let accessibility reviews slip while features piled up

I meant for TuckBuddy to be accessible from the start, and I reviewed for it early on. As features landed quickly, the reviews stopped. Then my partner, who runs larger text, opened the app and it fell apart. Text clipped inside tiles that couldn’t grow and forms were hidden. Most of it traced back to my custom look: fixed type, fixed tile sizes, and white text on pale glass.

Image to comeThe custom look at a large text size, with the clipped tiles and low-contrast text called out

I took the app back to stock iOS with teal accents and a few custom moments, and tiles gracefully fall back to lists when space is tight. Before launch I ran Apple’s accessibility audit, fixed the nearly 260 issues it flagged, and set the VoiceOver reading order on critical screens.

Image to comeBefore
Previous build
Image to comeAfter
1.0

Next time, I’ll require the bot crew to run accessibility audits as part of acceptance criteria.

I lost time on sharing before I understood how hard it was

I had no experience coding anything like sharing, or with CloudKit. The bot-workers’ research was light, and it lulled me into a false sense of confidence. I should have run more research before committing time to it.

I also didn’t account for how fragile my own test setup was. Testing sharing took two separate iCloud accounts, which wasn’t feasible in the time I’d given it. I shipped 1.0 single-user and moved sharing to the next build.

On my next project I will check my own understanding of challenging features and find ways to build in automated testing to watch for problems before wasted time piles up.

05Outcome

Where it landed

I made a useful thing that reflects my values and gets real jobs done. It is launched, usable, and useful in ways nothing else currently is. It’s too new for retention data, so the next build adds the signals that will tell me which action marks a real user. That build also adds multi-user sharing and sync over CloudKit, the paid tier that keeps the core app free.

How I designed and worked with the bot-workers is its own case study.

Working docs: requirements, roadmap, business notes, onboarding, habits and flywheels, engagement, activation.

See more case studies →