Integrating Generative AI into web applications often feels like a wild west of raw API calls, unstructured prompt strings, and security headaches. But building intelligent web apps doesn't require a degree in machine learning, just the right abstractions.

Enter Firebase Genkit and Firebase AI Logic: open-source frameworks designed to help web developers build, test, and deploy production-ready AI capabilities using familiar, full-stack patterns.

In this session, we'll walk through building a complete AI-driven feature from scratch. You’ll learn how to model structured AI flows, evaluate prompts locally using the Genkit Developer UI, and seamlessly stream responses back to the client. To show how framework-agnostic your backend AI logic can be, we'll look at integration patterns in both Angular and React.

Key Takeaways:

- How to structure type-safe AI workflows using Genkit Flows and Firebase

- Best practices for prompt engineering, structured outputs (JSON schemas), and local testing

- Clean client-side consumption patterns (Angular services/signals & React custom hooks)

- Security and rate-limiting patterns for user-facing AI endpoints