It’s dinner time and you are craving takeaway but you are determined to be healthy and cook at home. You only have a bunch of leeks in the fridge. Wait, there’s an app for that! OK Gemini, What’s for dinner?

This talk will document our journey in building a Kotlin multiplatform proof of concept app that helps you cook at home. You provide a photo of your dream dish and a list of ingredients you have. The app will will use multimodal Gemini AI to produce a recipe and a shopping list. We will explore what it takes to build the app and give some impressions of the experience. We will take the app for a spin and see what recipes it produces.

At the end of this talk the audience will get the following “takeaways”

- An idea what it takes to use Gemini AI with text and image in a multiplatform app
- A demo of the proof of concept app
- Explanations and gotchas we collected on the way
- An open source repo with the source

This talk is brand new and has not been accepted anywhere. It is suitable for intermediate Android with little to no AI experience. This will be a co-presentation of two speakers. It is a 40min talk.