Skip to content
Nomsters

Nomsters

As a first-time founder, I saw an opportunity to use AI to merge MyFitnessPal with Tamagotchi because sometimes we take care of our virtual pets better than we take care of ourselves.

Role
Founder: design, engineering, research
Dates
2024–2026
Focus
zero-to-onellmfluttergamesresearch
Nomsters: We Are What We Eat. A player and her pet look out over a lake of floating fruit and vegetables.

Build time

6 months → 2 months

Unity, 1 FTE + 4 PTE → Flutter + Cursor, 1 FTE + 1 adviser

Retention, D1 / D7

53% / 30%, vs a 26% / 15% health-app average

73-player camera prototype test, Oct 2024

AI-written code

20% → 80%

Nomsters 1.0 (Unity) vs 2.0 (Flutter)

Concept survey

90% excited, 82% would download

223 Gen Z and Millennial gamers, Apr 2024

After losing 30 pounds in 2023 with MyFitnessPal and MacroFactor, I wanted to turn the tedium of a caloric deficit into a hero's journey. Nomsters is a nutrition-based mobile game: snap a picture, and the meal becomes a food card you use to feed your pet, play minigames and complete quests.

The pitch

As gamers we live in worlds designed for engagement, so nutrition apps feel like homework, and most people quit them in their first week. Nomsters meets gamers where they already are: part Pokémon and Tamagotchi, part MyFitnessPal.

Nomsters: We Are What We Eat. Helping gamers live healthier lives through nutrition-based gameplay and virtual pets.
The problem: 3.4 billion gamers, more than 40% overweight or obese, underserved by last-gen apps and lessons.
The solution: harness a passion for gaming, and let players play with their food.
The product: a GenAI mobile game fueled by your nutrition, language-learning style gamification for food tracking.
Traction: concept surveys of 223 and 104 gamers, and a 73-player prototype test that beat the health-app average at D1 and D7.

Player Feedback

Before building the full game, I put the concept and then a rough prototype in front of real players through PlaytestCloud, in three rounds:

  • Apr 2024 · Concept survey, 223 Gen Z and Millennial gamers. 90% excited, 82% would download, 76% said they would play daily.
  • Oct 2024 · Re-survey, 104 players, after they saw the prototype trailer and the camera. Interest held and rose: excitement 97%, download 94%, and 84% saw the game as very aligned with promoting health.
  • Oct 24–31, 2024 · 7-day camera prototype playtest, 73 players on Android. D1 53% and D7 30% in the first cohort, 38% and 20% averaged across all start days. Players only had to play five minutes to be paid; 83% played past the first day.
Players want Nomsters: the April and October 2024 survey results, and the 73-player prototype playtest at D1 53.1% and D7 29.7%.

"The camera function was awesome. It was able to figure out it was coffee with milk and a ham and cheese sandwich." — playtester, Oct 2024

October 2024 re-survey of 104 players: excitement 90 to 97%, will download 82 to 94%, very aligned with promoting health 76 to 84%.
Art style sentiment: cartoon and illustrated rated highest for appeal, time spent and brand fit.
Gameplay sentiment: pet collecting and food trivia rated most exciting and most able to teach nutrition.
Audience expectations: players want a story and a game-y world, and still expect health outcomes.
Playtest takeaways: missing login and food-recognition accuracy were the top issues, and the actions taken on them.

A meal becomes a game object

A photo of a plate is recognised, mapped to a typed food card, added to your collection and synced. CV results were parsed into JSON, estimates grounded and revalidated against USDA Food Central (and other APIs), and would be processed into our game's food card system.

Goal was to get players to think of their diet as a TCG deck: each card has a mana cost, one mana is 100 calories, and a deck under 20 mana loses weight.

How it works: take a picture, track the food you eat, feed food to your pets, keep playing to learn about your nutrition.
The computer vision workflow: camera, food recognition, food card, collection, local cache and sync.

From struggling in Unity to vibing in Flutter

The Unity prototype took 6 months with a team of 5. After the solo pivot to full-AI, I rebuilt it in Flutter in just 8 weeks, with more of the game in it (story mode, collections, lesson plans, a second character) and web alongside iOS and Android. The code went from 18k lines of C# to 60k lines of Dart, and from 20% AI-written to 80%.

Struggling in Unity, built in 6 months, to vibing in Flutter, built in 2 months.
Nomsters 1.0 against 2.0: project length, features, front end, back end, code and team size.

From this study

Nomsters: We Are What We Eat. A player and her pet look out over a lake of floating fruit and vegetables.

A nutrition game where you play with your food

Nomsters