Building What You Need, When You Need It

What I Built

A wellness platform that tracks nutrition, alcohol consumption, sleep patterns, and personal transformation—all in one place. What started as a simple calorie counter has evolved into a comprehensive health tracking tool with AI-powered food recognition, a financial bridge to your retirement plan, and a privacy-first design that keeps your data on your device.

The platform now includes:

• Food tracking with quick entry for meals, protein, and calories
• Food image recognition — photograph a meal and the app identifies what’s on the plate and estimates nutritional content
• Nutrition label scanning with OCR — point your camera at any packaged food label and macros are extracted automatically
• FDA nutrition database search — find any packaged food by name and pull verified nutrition data
• Alcohol tracking with drink logging, calorie calculations, and weekly pattern insights
• The Real Cost of Drinking — projects your weekly-to-5-year alcohol spend in dollars and calories, plus empty-calorie body weight impact
• The Dividend — shows what redirecting your monthly drinking spend to retirement savings would do: years earlier you could retire, additional retirement runway, compounded corpus. Auto-syncs with your Reckon Wealth planner data if you’ve used it
• Sleep tracking with duration, quality, and trend analysis
• Interactive transformation journey documenting personal health milestones
• Privacy-first: all data stored locally in your browser, no account required

Try Health & Wellness →

Why I Built It

It started with a spreadsheet. I was tracking nutrition manually—functional but clunky. With AI-assisted development, I had a better option: just build what I needed.

The initial food tracker took a weekend. But as I started paying more attention to overall health, the tool grew naturally—first alcohol tracking, then sleep monitoring, then documenting the personal transformation that followed. Each feature was added iteratively, solving real needs as they emerged.

This is the power of AI-assisted development: not just building fast, but evolving fast. A tool that adapts as your understanding of the problem deepens.

How I Built It

Tools Used: Claude Code for development, React for the interface, Tailwind CSS for UI styling, Netlify Functions for server-side AI proxies, Web Speech API for text-to-speech, AI-powered image recognition and OCR for food logging, FDA FoodData Central API for nutrition data. Deployed on Netlify. All data stored locally in the browser — no backend database.

Phase 1: Food Tracker

I described what I wanted. Claude Code scaffolded the React app, implemented the form logic with Tailwind styling, set up local storage, and handled the math. The first version was live and useful within a weekend.

Timeline: One weekend—from “this spreadsheet is annoying” to production-quality app.

Phase 2: Alcohol & Sleep Tracking

As I started paying more attention to overall wellness patterns, I added two new tracking modules. The alcohol tracker logs drinks by type (beer, wine, liquor, cocktails, seltzers) with calorie calculations and cost estimates. The sleep tracker captures duration and quality on a five-point scale, with a seven-day bar chart showing trends at a glance.

Both modules use the same shared data layer and date navigation as the food tracker — proving that a well-architected initial build makes adding features almost trivial with AI assistance.

Timeline: A few hours per module, built across two sessions.

Phase 3: Image Recognition & OCR Food Logging

The biggest friction in food tracking has always been data entry. If logging a meal takes two minutes, you skip it. This phase tackled that head-on with two computer vision features that make logging nearly effortless.

Food Image Recognition — Take a photo of any meal and the app identifies what's on the plate, estimating calories and macros automatically. Useful for restaurant meals, home cooking, or anything where you don't have a label in hand.

Nutrition Label OCR — Point your camera at any packaged food label and optical character recognition extracts the nutrition data directly—calories, protein, carbs, fat—and pre-populates the log entry. No typing required.

Together, these turn logging from a data entry chore into a two-second photo. That changes the compliance dynamic entirely: people actually log when it's this fast.

Timeline: Built across two sessions, leveraging browser camera APIs and AI-powered image analysis.

Phase 4: The Real Cost of Drinking & The Dividend

The biggest insight from building the alcohol tracker wasn’t technical — it was behavioral. Tracking your drinks is one thing; seeing what they actually cost, in dollars, calories, and potential retirement years, is another.

The Real Cost of Drinking — After logging drinks, the tracker calculates your average daily consumption and projects the total cost forward: weekly, monthly, annually, and over five years. Alongside the dollar figure, it shows the calorie equivalent (and its body weight impact), and a “What If You Stopped?” panel showing how much you’d save in a year.

The Dividend — A single CTA bridges health and wealth. Click “See what this money could do for your retirement” and The Dividend appears. It takes your monthly drinking spend, pulls your retirement data from the Reckon Wealth planner via shared localStorage, and calculates: how many years earlier you could retire, how many additional years of retirement runway you’d gain, and the compounded corpus invested at 7% to age 65. If you haven’t used the planner, a quick manual form (age, current savings, monthly expenses) delivers the same calculation.

This is the most direct embodiment of what Reckon Well is about: the connection between how you live today and what your future looks like. The two pillars — Reckon Health and Reckon Wealth — are not separate tools that happen to share a brand. They share data.

Timeline: One session for The Real Cost of Drinking; two sessions for The Dividend, including the cross-product data bridge.

What I Learned

Architecture Insights

• Start simple, stay simple — Began with localStorage and never needed more. Each new module (alcohol, sleep, The Dividend) plugged into the same data layer without a backend. Don’t add infrastructure until you have to.
• Good initial architecture pays compound interest — Because the food tracker was built with clean component structure and a shared data layer, adding alcohol, sleep, and the cross-product Dividend feature was almost trivial. AI leverages existing patterns beautifully.
• Multiple modules, one codebase — Managing state across food, alcohol, and sleep tabs within a single React app taught real lessons about component composition and shared date navigation.
• Privacy-first architecture builds trust — Keeping all data in the browser (localStorage, no server) made the app simpler, faster, and more honest with users. Data that never leaves your device is data users trust.

Creative & UX Insights

• AI enables narrative experiences, not just CRUD apps — The My Story transformation journey combines SVG avatar animation, TTS voiceover, scroll-driven progression, and real-time data visualization into an interactive storytelling experience. This would have been a multi-sprint effort for a team; it was built in two sessions.
• Browser APIs are underutilized — The Web Speech API for text-to-speech turned static narrative text into an engaging voiced experience with minimal code. Natural pacing and sentence queuing required iteration, but the capability is built into every browser.
• Removing friction changes behavior — Image recognition and OCR didn't just make logging faster; they changed whether people log at all. A two-second photo beats a two-minute form every time.
• Mobile-first matters — Wellness tracking happens on your phone. Every feature was designed for thumb-friendly interaction first.

Development Insights

• Iterative expansion beats big-bang planning — Each phase (food, alcohol, sleep, My Story) was a self-contained addition. No grand upfront design. Ship, use, learn, expand.
• UI frameworks + AI = rapid polish — Tailwind CSS made every new module look cohesive and professional without custom CSS wrestling.
• Shared localStorage as a data bridge — The Dividend works because the health app and the Reckon Wealth planner use compatible localStorage keys. No API, no backend sync needed. The simplest architecture that could possibly work — and the connection it enables is the product’s most compelling feature.

Business Implications

This project—a wellness platform that evolved from a weekend food tracker into a multi-module platform with interactive storytelling—demonstrates how AI changes not just initial builds, but sustained product evolution:

1. Iteration Speed Compounds

The food tracker took a weekend. Adding alcohol and sleep tracking took a few hours each. The interactive My Story experience took two sessions. Each addition built on the last. This is the real unlock: AI doesn't just help you build fast—it helps you evolve fast. Products that would have required quarterly planning cycles can now grow weekly in response to real usage.

2. One Person Can Build a Platform, Not Just an App

This isn’t a single-feature tool anymore. It’s a multi-module wellness platform with food logging, AI image recognition, FDA nutrition search, alcohol tracking, sleep analysis, a real-cost drinking calculator, a retirement dividend projection, and narrative storytelling with TTS voiceover. Five years ago, this scope would have required a product manager, a backend developer, a frontend developer, and a designer. With AI: one person, built iteratively across sessions.

3. Creative Experiences Are No Longer a Luxury

The My Story feature isn't a typical CRUD form—it's an interactive, narrated journey with an evolving illustrated avatar, animated health metrics, and text-to-speech voiceover. This kind of narrative UX used to require specialized creative development skills. AI assistance makes it accessible to anyone with a vision for what the experience should feel like.

4. The Build vs. Buy Calculus Changed

Commercial wellness apps do 10x more than you need, charge subscriptions, and own your data. AI-assisted development creates a third option: build exactly what you need, keep your data, and evolve it as your needs change. For internal tools, niche use cases, or personal projects, the economics shifted permanently.

5. Good Architecture Enables Compounding Returns

Because the initial food tracker was well-structured (clean React components, shared data layer, consistent date navigation), every subsequent feature slotted in cleanly. The lesson: AI makes initial builds fast, but thoughtful architecture makes expansion almost free. This is where experienced product thinking still matters enormously.

The Pattern

This wellness platform follows the same pattern as all my AI-assisted projects—but extends it into something more powerful:

1. Real problem (not a tutorial or demo)
2. Start simple (localStorage first, cloud later)
3. Ship and use it (daily usage reveals what's actually needed next)
4. Expand iteratively (food → alcohol → sleep → narrative experience)
5. Push creative boundaries (TTS voiceover, animated avatars, interactive storytelling)
6. Evidence-based learning (what worked, what didn't, what it means for teams)

The point isn’t “look at my wellness app.” The point is: a weekend project can evolve into a multi-module platform that bridges health and wealth, connects two separate products through shared data, and delivers insights no commercial app offers—all built by one person, iteratively, using AI. That’s the shift.