Consumer Mobile App

Lattice

A mobile-first inventory app for crafters, shipped to the App Store in 7 months using AI-assisted development.

Type

Consumer Mobile App

Platform

iOS (App Store)

Role

Solo Founder, Designer and Developer

Timeline

September 2025 to April 2026

Lattice inventory list
Lattice project planning
Lattice pattern library

0.0%

App Store Conversion Rate, First 5 Days

7 mo

Concept to App Store, Solo

0

Core Feature Areas Shipped at Launch

tl;dr

Crafters shop blind. I built the fix.

Hobbyist crafters have no reliable mobile way to track their supplies. Spreadsheets break on mobile. Warehouse tools are overkill. Nothing on the market was built for someone standing in a craft store trying to remember what they already own.

Lattice is a mobile-first inventory system covering supply tracking, project planning, and pattern organization. Shipped solo in 7 months using AI-assisted development.

Problem

The gap became obvious in context.

The User

Hobbyist crafters and multi-medium artists who practice two to four crafts seriously, as hobbies or small side businesses. They buy supplies frequently, store them across multiple locations, and consistently lose track of what they have.

  1. Duplicate purchases

    Crafters regularly shop without knowing what they already own. Stained glass sheets run $30 to $60 each.

    A hand holding a sheet of black iridescent stained glass in a home craft space, with storage bins and a craft store bag in the background.

    One sheet of stained glass. Crafters store dozens across multiple locations.

  2. Stalled projects

    Missing a single material mid-build stops everything.

  3. No mobile reference

    No mobile reference while shopping. No record of what is at home.

  4. Broken workarounds

    Spreadsheets that don’t work on mobile, camera roll photos with no metadata, or nothing at all.

    A phone camera roll full of photos of stained glass sheets held up by hand, with no names, brands, or prices attached.

    The old workaround: a camera roll of glass photos with no names, brands, or prices.

Why Existing Apps Failed

I looked at 12 competitor apps before writing a line of code. Consumer apps like Easy Stash, MakerLog, and Krafts Inventory were built for single crafts, plagued by trust and paywall issues, or abandoned by their developers. Business inventory tools like Sortly were designed for warehouses: too complex, too expensive, and missing features crafters actually need.

Easy Stash
Easy Stash
MakerLog
MakerLog
Krafts Inventory
Krafts Inventory
Sortly
Sortly

Apps reviewed during competitive analysis

The gap: nothing mobile-first, craft-agnostic, and priced for hobbyists.

Solution

One stash, connected to every project.

Supplies, projects, and patterns work on their own and are more useful together. Every UX decision was made for someone standing in a store, not sitting at a desk.

  1. Stash

    Document supplies with photos, quantities, and custom categories. Search and filter to find items without scrolling.

    Lattice inventory: stained glass supplies with photos, quantities, colors, brands, and storage tags.

    Each supply with a photo, color, brand, and where it is stored.

  2. Item details

    Quantity, brand, color, storage location, and tags for every supply, with its photo.

    Item details for Black Iridescent glass: a photo, quantity, brand, color, storage location, and tags.

    Item detail view: every field crafters actually need, nothing they don't.

  3. Projects

    Create projects, track progress, and build task lists.

    Project details for Sunrise Panel for Jacob: status, progress, start date, and a task list.

    A project with its status, progress, and tasks.

  4. Connected

    Each project keeps the materials it needs and the pattern it follows together.

    A project showing its linked inventory items, Black Iridescent, Amber, and Amazon green glass, plus its primary pattern.

    The project now holds its glass and its pattern in one place.

  5. Patterns

    Upload and organize craft patterns, PDFs or images. Tag by difficulty, type, and material. Favorite patterns for quick access.

    Lattice pattern library with suncatcher patterns tagged by difficulty, material, and type.

    Patterns tagged by difficulty, type, and material.

Adding an item, straight from the camera. Sped up.

Process

Designing directly in code to accelerate iteration.

As a solo builder, I replaced traditional design-to-handoff workflows with a tighter loop: decision, implementation, validation. Building in a live environment forced earlier confrontation with real constraints like scroll behavior, state management, and load times, and produced better decisions faster than static Figma screens alone.

Phase 1

Scope and strategy

Rather than building everything I could imagine, I locked the MVP to four areas: supply tracking, project management, pattern organization, and a freemium subscription model. Every scoping decision was deliberate.

  1. 40 free items, not 75

    Enough to demonstrate value without eliminating the reason to upgrade.

    Side by side: an early Pencil.dev pricing layout with 50 free items and Plus at $4.99 a month, and the live Lattice subscription screen showing the free plan's limits: 40 inventory items, 3 active projects, 10 saved patterns, and 1 photo per item.

    The free plan as shipped: 40 items, 3 active projects, 10 patterns, 1 photo per item.

  2. $2.99 per month, not $4.99

    Lower friction for hobbyists skeptical of yet another subscription.

    Side by side: an early Pencil.dev pricing layout with Plus at $4.99 a month, and the live Lattice Plus screen offering $24.99 a year or $2.99 a month.

    The early layout priced Plus at $4.99. It shipped at $2.99 a month, or $24.99 a year.

  3. All crafts, not yarn-first

    Competitors narrowed to one craft. I built craft-agnostic from the start.

    Side by side: an early Pencil.dev inventory layout filled with fabric and yarn, and the shipped Lattice inventory filled with stained glass supplies.

    Sketched with fabric and yarn, shipped with stained glass. The same inventory works for both.

Phase 2

AI as an execution layer, not a decision-maker

I used AI-assisted development tools (Claude Code, Cursor) to implement product decisions, not generate them. I defined what to build, how it should behave, and what tradeoffs to make. AI accelerated execution by translating those decisions into working code. That division of labor is what made solo product development feasible without compromising product quality.

One workflow refinement: I initially prompted Claude Code with written descriptions of the UI I wanted. Results were inconsistent. Switching to Pencil.dev, a tool that lets you sketch rough layouts on an iPad and feed them directly to Claude Code as visual references, significantly tightened the gap between what I described and what got built.

Pencil.dev canvas with layouts for the pricing page, inventory list, add item form, inventory dashboard, and projects screen.
Pencil.dev layouts fed directly to Claude Code: the sketch-to-code workflow that replaced Figma handoff.

Phase 3

App Store submission as a design problem

Preparing for App Store submission involved more design work than I expected. Screenshots, preview copy, category selection, and pricing configuration all required the same intentionality as the app itself.

The live App Store product page.

Outcome

Early signal: strong product-market clarity.

In the first five days, 18.2% of App Store product page views converted to downloads, against a typical 3 to 5% for utility apps, with no paid acquisition behind it.

The sample is still small, and I treat it that way. What the early rate suggests is alignment between the problem, the positioning, and the product page. That is the signal the next iteration builds on.

Product page conversion, first 5 days

Lattice18.2%
Typical utility apps3 to 5%
App Store product page views that became downloads, with no paid acquisition. Small sample.

What I learned

↑ Process

Building in code changed how I make design decisions.

Designing in a live environment forced earlier confrontation with real constraints: scroll behavior, state management, load times. I made better decisions faster than I would have from static Figma screens alone.

↑ Scope

Scope discipline is what makes solo shipping possible.

Every feature I did not build in v1 was a deliberate choice. The freemium model, the item limits, the iOS-first strategy. All were cuts that let me ship something complete rather than something sprawling.

↑ Launch

The App Store listing is a product, not a form.

Screenshots, preview copy, and category selection drove an 18.2% conversion rate in week one. That required the same design thinking as the app itself.

Every decision about what to build, what to cut, how to price, and what the experience should feel like was mine. Claude Code handled the code that implemented those decisions. That division of labor is what made the project feasible as a solo non-developer, and it is a workflow I would use again.

A note on how this was built

Lattice is available on the iOS App Store at craftwithlattice.com

What's Next

Distribution comes next. Then iteration.

The app is built. The gap now is getting it in front of crafters. Community-based distribution, craft Facebook groups and Instagram, is the current focus. The next product iteration will be grounded in real user behavior data rather than pre-launch assumptions.

Lattice dashboard showing active projects, inventory count, and a stash value of $15.49.

Shipped since launch · v1.0.6

Know what your crafts really cost.

Log what you paid when you add an item. Restocks are kept as separate purchases, and the dashboard shows the value of everything in your stash.

Real user research

I used AI to pressure-test assumptions pre-launch. Now there are real users. The next design decisions will be grounded in actual behavior data.

Android and barcode scanning

Android is the next platform. Barcode scanning is the most requested feature from craft communities: add an item by scanning a product barcode instead of entering it manually.

More Case Studies