MainMeal - adapt meals for every person, age, and condition

Families all eat differently. MainMeal is AI that adapts recipes to each person’s health needs—low salt, baby-safe, low purine and more—so one kitchen can cook right for every body.

Comment

Inspiration

MainMeal was inspired by two areas of our lives.
From our healthcare background, we saw how people with chronic conditions like diabetes and high uric acid struggle to choose meals that support their health. Many families are unsure how to cook one dish that everyone can safely enjoy.
As parents of a toddler, we also deal with safety rules, picky eating, and changing tastes. Managing the needs of multiple generations in one kitchen often feels overwhelming.
This led us to ask a single question: Can any recipe be adapted for everyone at the same table?

What it does

MainMeal takes any recipe and automatically adapts it for each family member.
Users paste a recipe, and the system returns:

  • A base description of the dish
  • Per person verdicts (ok, needs adaptation, or not recommended)
  • Clear explanations
  • Personalised adaptations
  • Safe options for toddlers and people with dietary conditions

It supports diverse needs such as picky eating, diabetes, high uric acid, and baby-safe requirements.

How we built it

We designed structured family profiles containing roles, hard restrictions, and simple health indicators. These profiles are stored within the app and used during recipe processing.

We built a clean Next.js frontend and used Cursor to write, troubleshoot, and generate code while learning development for the first time.
All project files were managed with GitHub, which we learned from scratch during the hackathon.
Deployment was done through Vercel, allowing us to share the live prototype instantly.

For the AI logic, we integrated Anthropic models to analyse recipes, check constraints, and return structured JSON with adaptations for each person.
We tested recipes such as tiramisu, fried rice, and chicken stir fry until the outputs became consistent, safe, and practical.

To produce a polished demo, we used multiple creative tools:

  • Cursor for rapid code iteration
  • Gemini for narration and script ideas
  • CapCut (AI voice generator) for clear and consistent voiceovers
  • Canva for branded visuals
  • ChatGPT for copywriting, editing, and structuring the story
    The final demo was created through screen capture and edited into a complete video.

Challenges we ran into

Starting with zero technical experience was our biggest challenge.
We had never set up GitHub accounts, never used Git Bash, and had never run a project locally. We learned basic commands, installation steps, and how to interpret error messages under time pressure.

We also had to learn version control, repository management, pushing commits, and connecting GitHub to Vercel for deployment.

Understanding how a full stack project works was another challenge. We had to figure out how the frontend, backend, and AI logic communicate, and how to debug issues when things broke.

Learning to use multiple new tools at once (Next.js, GitHub, Vercel, Cursor, Anthropic, Gemini, CapCut, Canva) required patience and teamwork.
To stay focused, we kept reminding ourselves of one goal: make the demo reliable and easy to understand.

Accomplishments that we're proud of

  • Building a working AI system despite starting with no technical background
  • Learning GitHub, Git Bash, and running real code for the first time
  • Creating a functional recipe adaptation tool that handles multiple dietary needs
  • Deploying our first web application using modern frameworks
  • Producing a polished, professional demo video
  • Collaborating effectively under tight time limits

What we learned

We learned how to build an end‑to‑end application from scratch, including version control, frontend development, debugging, and prompt design.
We also learned how dietary needs can be represented in a structured format so that AI can adapt recipes accurately.

What's next for MainMeal — adapt meals for every person, age, and condition

We plan to expand MainMeal with features such as:

  • More dietary and medical profiles
  • Adjustable portion sizes and nutrient summaries
  • Grocery list generation
  • Integration with recipe websites
  • User accounts for saving profiles and history
  • Allergen awareness and cultural dietary support

Our long-term goal is to help families cook confidently, safely, and inclusively, no matter how many different needs they manage at mealtime.

  1. Subscription Ecosystem We plan to build a scalable subscription model that grows with households over time.
  • Free Tier

  • Entry-level access with basic recipe adaptation and a limited number of family profiles

  • Premium Tier (RM12–RM28/month)

  • Designed for families managing multiple dietary needs, offering:

  • Unlimited profiles, advanced medical indicators, nutrient summaries, adjustable portions, grocery lists, browser extension support, and priority processing.

  • Future Positioning: “A single subscription that keeps every family member safe, included, and confident at mealtime.”

  1. Grocery Advertising & Retail Partnerships
  • As MainMeal evolves, we will open the platform to grocers and FMCG brands who want to reach high-intent home cooks.
  • Planned advertising formats include: -- Sponsored ingredient substitutions -- “Brand X low-sodium soy sauce — recommended for heart-safe cooking.”
  • Featured grocery items within adapted recipes and weekly meal plans
  • Retail campaigns tied to festive menus and seasonal cooking trends
  • Future integrations may support dynamic pricing, real-time promotions, and supermarket-specific supply preferences.

Built With

  • canva
  • capcut
  • chatgpt
  • cursor
  • elevenlabs
  • geminiapi
  • gitbash
  • github
  • googleflow
  • nextjs
Share this project:

Updates