Verdu: Your AI Nutrition Coach

Stop guessing and start achieving! Verdu is the all-in-one AI nutrition coach that makes healthy eating intuitive. Track macros, plan meals, and get smart grocery lists.

Comment

Background & inspiration

A few years after meeting in Costa Rica, we finally seized the opportunity to collaborate. Inspired by the latest advancements in mobile development AI tools, we began brainstorming at Damia's jungle house. Our conclusion was simple: most people want to eat healthier but need more than just calorie counting—the standard for most nutrition apps. They need a true nutrition coach that teaches them how the food they eat impacts their specific goals. With the power of AI, we realized this vision was finally achievable.

The idea for Verdu was born: an all-in-one nutrition coach that guides the user through the entire journey to better health. This meant integrating meal tracking and analysis, smart shopping list and pantry management, personalized meal ideas, and continuous coaching support.

What it does

After a quick onboarding, Verdu calculates your daily and weekly macro goals. From there, it actively steers your journey with a fast, efficient, and intelligent set of features:

  • Goal-Aligned Meal Ideas: When it's time to eat, Verdu provides meal ideas perfectly aligned with your current macro targets, prioritizing ingredients you already have in stock.
  • Smart Eating Out: Verdu’s restaurant menu picker helps you quickly identify goal-friendly options.
  • Dish Optimization: Simple in-meal tools offer easy swaps and portion tweaks to keep any dish on target.
  • Continuous Learning: Verdu teaches you with every interaction, highlighting what's good and bad about a meal and suggesting specific ways to improve it for next time.
  • Kitchen Automation: Features like automated grocery list creation and inventory management simplify meal prep and eliminate food waste.
  • Transparent Progress Tracking: Beyond standard charts, Verdu provides a simple, consolidated view that connects calorie/macro intake directly to weight changes, ensuring small progress or minor misses are always visible and actionable.

How we built it

We built Verdu using Flutter and leveraged Claude Code as a tightly directed AI pair programmer. The foundational step was creating a robust Design System that defined our colors, typography, spacing, motion, and a comprehensive, reusable component library.

In our CLAUDE.md, we enforced strict rules:

  • Always reuse Design System components. If a new component or pattern is needed, generate it once for immediate reuse.
  • Every screen must include micro-animations and haptic feedback.
  • Adhere strictly to SOLID, KISS, and DRY coding design principles.
  • Build all screens using the BLOC pattern, repositories, and reactive streams.

These constraints kept the AI output consistent, shippable, and resulted in a surprisingly high level of design excellence for a first-pass MVP. Approximately 98% of the code was generated by Claude Code with minimal supervision, 1.9% required heavy guidance, and only 0.1% was hand-written. We also integrated RevenueCat for subscription management.

Challenges

Our biggest challenge has been ensuring the AI-generated code consistently adheres to the specific constraints of our custom Design System. Looking ahead, we anticipate challenges in refactoring the codebase, integrating new, key AI features, and most importantly, driving product adoption through effective marketing.

Accomplishments

  • We successfully built the entire MVP in under two months.
  • The app works: Early testing with friends and family has been extremely positive, including one user who has already lost 15lbs!
  • We're proud of the final product's design quality, especially considering no dedicated designer was involved.

What we learned

  • Knowledge Drip: Trickling small, relevant pieces of nutritional knowledge to users with each interaction significantly increases feature stickiness and long-term engagement.
  • Nutrition Cycle Management is Key: Helping users manage the most annoying and time-consuming parts of the nutrition journey—like shopping and inventory—delivers the highest perceived value.
  • Solid Rails: Establishing a strict Design System and explicit AI coding rules keeps large-scale generative development highly consistent and on track.

What’s next for Verdu

Our immediate focus is singular: achieving Product-Market Fit. This means an intensive cycle of Marketing, Exposure, Iteration, Feedback, and more Marketing!

Built With

Share this project:

Updates