Getting started with Genkit

Genkit is an AI integration framework designed for app developers. It allows you to easily integrate powerful AI capabilities into your apps with familiar patterns and paradigms. Join @PeterFriese to learn what Genkit is, get to know the key concepts, and see how to build an AI-powered app along the way.

Chapters:
0:00 - Cold open
1:05 - Installation and getting around the DevUI
3:00 - Introduction to Genkit Flows
5:36 - Using open source models via Ollama
6:23 - Implementing your first Genkit Flow
10:55 - Calling Genkit flows from a Firestore trigger
11:30 - Inspecting traces
13:37 - Implement RAG-based Q&A
16:37 - Compute vector embeddings and store them in devLocalIndexer
18:22 - Retrieve matching documents using devLocalVectorstore
20:04 - Demo
21:09 - Storing Vector Embeddings in Firestore
24:19 - Prompt management with dotprompt
28:18 - Deployment
29:45 - Next steps

Resources:

#FirebaseFundamentals #Firebase #Genkit #ai

Speaker: Peter Friese
Products Mentioned: Firebase, Firebase Genkit

[antonkulakovcom]

This seems to be useful for my application. Thanks for this video!

[d.g.]

This right here is some DX as known from firebase. Good to see how it picks up the already established metaphors around AI development and simplifies it to a level where it feels like getting started with AI development in general, is like a no-brainer.

[rebarius]

Very good introduction and explanation ❤ thanks ☺️👏🏻

[IowaKnight]

Loved the example! Can you do a follow up with other types of vector store like BigQuery for vectors and Firestore for documents for example!

[周旭虎]

Why the GENKIT is written in Javascript? I really want to develop AI apps with Firebase in Python. Because I only can do pyhton.