Observe
Capture food, water, medication, symptoms, weight, behavior, and records.
Project 01 · AI-Native iOS Product
Floofy
An iOS health journal for families caring for cats with chronic conditions.
I independently took Floofy from research to App Store launch, using AI across product design, development, testing, analytics, and marketing.
500+ downloads · 5% paid conversion · Live since May 2026
Role — Founder + Product Manager
Scope — Research, strategy, UX, development, launch, growth
Built with — Claude Code, Figma, SwiftUI, Supabase, Cloudflare
Download Floofy on the App Store ↗
500+App Store downloads
5%Paid conversion
22KConcept video views
App Store figures updated July 2026 · concept reach shown separately from adoption
The Problem
The problem became clear in the exam room, but it began much earlier at home.
Caring for a cat with a chronic condition means making small decisions every day: what to feed, whether a symptom is unusual, whether a medication is helping, and when a change is serious enough to contact the vet. The information needed to make those decisions is often scattered across memory, notes, food labels, lab reports, photos, and disconnected online searches.
Core job to be done
The real job was not simply “track my pet.” It was: “Help me take better care of my cat every day, understand when something is changing, and feel prepared when I need veterinary support.”
Primary user: an owner actively managing a diagnosed chronic condition with recurring vet visits. Secondary user: a highly attentive owner who may benefit from the same tools, though I optimized for chronic care because repeated decisions make a continuous history worth the effort.
AI-assisted, human-verified research
I used AI to identify recurring patterns across public chronic-cat communities, then reviewed the original discussions before treating them as product evidence.
I used AI to identify recurring patterns across public chronic-cat communities, then reviewed the original discussions before treating them as product evidence.
- Reconstructing meals and symptoms before visits → structured logs and a vet-ready report.
- Searching food labels repeatedly → a condition-aware food database.
- Starting isolated AI conversations → record review attached to saved context.
User Journey
A continuous home-care system.
Floofy turns fragmented daily observations into a continuous care history, helping owners understand their cat’s needs at home and collaborate more effectively with their vet.
Understand
See patterns and compare changes against the cat’s normal baseline.
Act
Make more informed daily decisions about feeding, medication, monitoring, and when to seek help.
Share
Bring a reliable longitudinal history to the vet when professional care is needed.
The Pivotal Decision
Public testing revealed that generic tracking was not valuable enough.
The first concept never reached the App Store. I recorded a walkthrough, shared it publicly, and used the response to test where logging created enough recurring value.
- 01 · Prototype
Start broad
I built a general pet-health tracker for owners who wanted one place for daily care.
- 02 · Share
Test the story, not the release
I recorded a product walkthrough and posted it on Chinese social media before investing in a production launch.
- 03 · Listen
Separate reach from evidence
The video reached 22,420 views. Qualitative feedback—not the view count itself—showed that generic logging felt like work without enough recurring value.
- 04 · Refocus
Build for chronic care
The strongest need came from families already managing consequential, repeated care without a coherent history.
What changed: the feedback did not invalidate tracking. It showed when tracking becomes worth the effort — when care is recurring, decisions are consequential, and history needs to travel to the vet.
The Product
One care loop, made visible.
Four connected capabilities help owners observe, understand, act, and share — carrying context from daily care into the next decision.
01
Log in under 10 seconds
Condition-filtered shortcuts reduce food, water, medication, symptoms, and weight to a few focused taps.
02
Compare food against this cat’s nutritional priorities
Search more than 900 foods and review them against condition-specific nutritional priorities.
03
Understand a record without losing its context
Floofy extracts important values, explains them in plain language, supports follow-up questions, and keeps the original record available for future comparison.
04
Turn logs into a vet-ready record
Floofy automatically organizes daily logs into a shareable history, so owners do not have to reconstruct care from memory.
Supporting utility: reminders for medication, fluids, weigh-ins, and recurring care.Completes the loop — not the core differentiator
DownloadView on the App Store ↗Live public listing
StatusLive production appiOS · shipped May 2026
AI-Native Execution
How I used AI across the product lifecycle.
Discover
- Synthesized recurring patterns from community research.
- Organized interview findings and hypotheses.
- Identified areas requiring manual source verification.
Design
- Used Claude to generate multiple UI directions from product requirements.
- Refined information architecture, flows, and interaction details in Figma.
- Created a reusable design system for typography, components, spacing, and states.
- Evaluated generated concepts against accessibility, comprehension, and care urgency.
Build
- Broke features into scoped implementation tasks with acceptance criteria.
- Used Claude Code to develop SwiftUI interfaces and application logic.
- Debugged integration, state, sync, payments, and edge cases.
- Maintained architectural decisions across local storage, cloud sync, and secured AI requests.
Measure
- Defined analytics events and built the product dashboard.
- Used AI to accelerate event-query creation and dashboard implementation.
- Tracked activation, food search, report creation, subscription conversion, and retention.
Launch and grow
- Prepared App Store content and screenshots.
- Developed launch messaging and social content.
- Used feedback and usage data to prioritize iterations.
Example 01 · Product judgment
Show less, based on the cat’s condition
- Initial AI approach
- The AI-generated PRD put every quick-log action on every owner’s homepage: food, toilet, vomiting, symptoms, toothbrushing, breathing rate, and more.
- Problem I identified
- Different chronic conditions create different daily priorities. A universal dashboard increased mental load and made the most relevant actions harder to find.
- My decision
- I predefined condition-specific templates. Kidney Issues surfaces food, weight, urine, and toilet; IBD surfaces toilet, weight, and food. An information control explains that owners can still choose what to show or hide.
- Result
- A more focused homepage that feels tailored to the cat’s needs without removing owner control.
Example 02 · Trust-boundary judgment
Gemini extracts. The app converts.
- Product problem
- Adding a custom food requires about 12 nutrition values, but package labels use as-fed percentages and energy units while Floofy’s suitability engine needs normalized values such as mg per 1,000 kcal.
- Tempting shortcut
- Ask Gemini to read the label and calculate every normalized value in one prompt.
- Risk I rejected
- A plausible arithmetic or unit-conversion error could silently feed a confident suitability rating for a medical-adjacent decision.
- My decision
- Gemini returns only the values and units printed on the label. A pure, unit-tested Swift converter resolves energy density and performs every derived calculation through a deterministic fallback order.
- Result
- Extraction errors remain visible in an editable form, calculations stay testable, and the UI identifies when energy is estimated rather than presenting it as fact.
LLM arithmetic never feeds suitability ratings.
SwiftUI + SwiftData
Native views, app services, and an on-device source of truth for fast, offline-capable logging.
Supabase
Authentication, Postgres with row-level security, file storage, and background synchronization.
Cloudflare Workers
Keeps model credentials off-device while routing and constraining AI requests.
Gemini
Supports record review, lab OCR, contextual conversations, and English/Chinese translation.
PostHog
Tracks product events, activation, feature use, conversion, and retention.
StoreKit 2
Handles subscriptions, trials, entitlement state, and localized App Store pricing.
Why these technical decisions mattered.
Local-first data
Logging stays fast and available with poor connectivity.
Background sync
Daily care does not pause while cloud data catches up.
Secured AI gateway
Model credentials remain off-device and requests can be controlled centrally.
Native payments
Subscription access fits the App Store experience users already trust.
Outcomes + Learning
What early traction revealed.
Results
500+
Downloads
Real adoption after launching in May 2026.
5%
Paid conversion
A more meaningful early signal than the absolute proceeds.
22K
Concept views
Reach that produced qualitative feedback before the production build.
What I learned
Visible pain makes a stronger acquisition hook
Nutrition posts for cats with chronic conditions already attract likes and saves because owners are actively searching for what they should feed. Floofy turns that scattered demand into one condition-aware food resource.
AI needs a defined stopping point
AI is valuable for finding patterns and reading messy inputs, but consequential calculations need deterministic, testable code. The product became more trustworthy when I designed that boundary explicitly.
Acquisition and retention are different jobs
Food information creates immediate value and brings owners in. The next product challenge is connecting that discovery to a saved cat profile, repeated care logging, and a longitudinal history that becomes more useful over time.
★★★★★
“Beautiful and very useful cat care app. I love the colors and clean design, and it makes tracking food, weight, symptoms, medications, and vet records so much easier. The food database are really helpful.”
What I Would Do Next
Turn discovery into a lasting care habit.
Improve activation
Deliver value before asking a new user to establish a logging habit.
Connect discovery to retention
Use food search as the entry point, then guide users toward a cat profile and care history.
Increase report completion
Show report readiness and progressively collect the information needed for a useful summary.
Strengthen trust
Make nutritional sources, medical boundaries, and condition-rating logic more transparent.
Thanks for reading.