Building MacrosAI from scratch
Overview
MacrosAI is a mobile app designed to simplify calorie tracking and support wellness goals. What sets MacrosAI apart is its unmatched versatility and ease of use.
By harnessing advanced AI and linking to the USDA food database, it enables users to quickly log any food item—from snacks to multi-ingredient meals—through a photo or a brief description.
This seamless integration of technology ensures accurate tracking in just a few seconds, making it easier than ever to stay on track with dietary goals.
Role
Co-founder & CEO
Platform
iOS
Deliverables
Discovery, customer management, product design and development, operations and front-end development
Execution
Problem
Today’s calorie tracking apps, despite offering barcode scanning, search, and manual entry, only facilitate quick and accurate logging for single-use snacks, like a Snickers bar, or standard items, such as an apple. However, logging meals beyond these simple items necessitates manual entry of portion sizes and ingredient weights by the users.
This pain point is compounded by the need to log meals multiple times daily, turning the process into a tedious and time-consuming task. Consequently, this leads some customers to abandon their calorie counting efforts altogether.
Initial hypothesis
As a fitness enthusiast, I’ve tried various apps like MyFitnessPal and Cronometer to track my daily calorie and macros intake. However, my engagement with these apps gradually decreased and eventually stopped.
The reason? It took too much time to constantly weigh, measure, and manually log every ingredient for home-cooked meals and takeouts.
I have then formulated my initial hypothesis:
Many individuals aim to track their calories to meet their wellness goals but find the process tedious and time-consuming, leading to inconsistent logging and eventually to quitting.
There’s a willingness among this group to trade a slight decrease in calorie tracking accuracy for a more streamlined and expedient meal logging experience.
Research
Solution formulation
At that time I have already been a Chat-GPT user for both work and personal use.
When Open AI released a new Vision capability, that’s when it clicked – I could use the power of Large Language Models to help people understand how much they are eating and what nutritional value they are getting from it.
I started experimenting with ChatGPT, logging my own meals and tweaking the prompt as my understanding of what I needed evolved.
Solution validation
Since I knew it was technically possible to log complex meals using a single image—by that time, OpenAI had already added Vision to their list of available APIs—I began researching the early adopter space to see if I could validate my solution with them.
I found several Facebook groups and Reddit sub-forums where people were asking for calorie advice. There, I started posting replies, attaching a screenshot from an app that didn’t exist yet, to see if they would be interested in signing up as beta testers.
To accomplish this, I manually ran their images through the OpenAI API and then constructed the UI in Figma.
Research outcomes
The feedback from early adopters was overwhelmingly positive, garnering hundreds of upvotes, comments, and requests from the community eager to join the waitlist for MacrosAI.
Crucially, this engagement led to in-depth conversations with potential users about their wellness and dietary journeys, shedding light on their objectives and providing valuable insights. These discussions were instrumental in shaping the product roadmap for MacrosAI, offering a wealth of user feedback on desired features and functionalities that could enhance the app’s utility and user experience.
This direct user interaction has been pivotal in aligning the development efforts with the real needs and preferences of my target audience, ensuring MacrosAI is finely tuned to serve its users effectively.
Building the MVP
After gaining a solid grasp on what the MVP for MacrosAI would entail, I teamed up with a technical co-founder to bring the concept to life.
My role centered around product development, design, analytics, prompt engineering for the OpenAI API integration, and tackling the front-end work using React Native. Meanwhile, my co-founder focused on the back-emd of our operation: the back-end development, cloud infrastructure and security.
This experience was highly educational, broadening my grasp on app development and API intricacies, thus adding a substantial layer of technical expertise to my product management skills.
Outcomes
As of May 2024, the app is fully functional on TestFlight, with over 100 daily active users and a growing waitlist of hundreds. We’ve maintained a 48% retention rate over a 30-day period, showcasing the app’s engaging user experience.
Currently, we’re concentrating on adding a monetization strategy and preparing for the official launch on the App Store.
