AI Meal Planner

AI MealPlanner
Meal Plan - Example
Meal Plan - Shopping list (1)
Meal Plan - Shopping list (2)
Recipe

Inspiration

Life in a busy household with a family of 5 :) - it was actually my wife's idea!

What it does

AI Meal Planner creates a personalised weekly meal plan for you and your family, tailored to your preferences and dietary needs. It suggests lunch and dinner recipes with clear ingredient lists and step-by-step instructions.

You’ll also get a smart shopping list that combines all the ingredients you’ll need for the week — making grocery planning simple and efficient. In the future, we aim to integrate with supermarket APIs to enable one-click ordering straight from the app.

Meal plans can be shared with others, and if you choose to log in, your plans are saved and accessible via a private shareable link. You can also add meals directly to your calendar, so each day’s recipe (with ingredients and instructions) is right there when you need it — even on the go.

How we built it

All built using Bolt.new and Supabase (Edge functions, Auth, DB).

Challenges we ran into

The idea was having a one-click shopping list to grocery cart, and being able to easily buy all ingredients you'd need for your week, however I couldn't find any supermarket in the UK offering such APIs.
Several headaches with race conditions between generating recipes and storing them on localStorage first and DB later 🤯 (all good in the end!).
When using Supabase magic link auth, some unexpected side-effects of leaving tabs open in the background with active user auth subscriptions.
Letting Bolt.new loose on the codebase was sometimes a bit scary, particularly towards the end. :)

Accomplishments that we're proud of

The variety of recipes generated, from ingredients to cooking styles and dietary requirements. You can even tell it to generate a pirate themed meal plan!
Working for both logged in and anonymous users.
DB persistence for logged in users, which allows sharing of meal plans.
The resilience of the app - if user refreshes mid meal-plan generation, the app is able to resume work where it left off. If user logs in after a meal plan is generated, it gets persisted to the database and saved into the users plans.
Building something useful, end to end (with very little coding).

What we learned

That we can ship a pretty much production grade app with very little coding!!

What's next for AI Meal Planner

I'd love to bring the one-click cart generation and order from supermarkets to live but unfortunately it's dependent on having access to the right APIs - maybe getting the app in front of them would be a good starting point.