Firebase is known as a backend as a service (BaaS) for mobile and web developers, and its realtime databases. But it also included a number of AI features that you can use to implement exciting new use cases for your apps.

In this talk, I will demonstrate how to implement an AI-powered meal planner using Firebase, bringing together Firestore, Gemini, Genkit, Cloud Storage, and SwiftUI.

In addition to walking you through plenty of code samples, I will explain the key concepts that underpin the technology powering the app. There is also going to be a good helping of live demos, which will include pictures of delicious recipes, so better make sure you’re not hungry when attending this talk.

You will learn

- How to securely call LLMs like Gemini and Imagen from mobile apps, without leaking your API keys
- How to use multimodal prompts
- How to generate text and images
- How to generate structured data
- How to use vector embeddings to implement semantic search
- How to use Retrieval Augmented Generation to tap into the user’s data stored in the app
- How to monitor your AI features and keep an eye on the number of tokens consumed and produced