googlecloudplatform/generative-ai

★ 16,836Jupyter NotebookAudience · developerComplexity · 2/5Setup · moderate

Mindmap

mindmap
  root((Google Gen AI))
    Topics
      Gemini models
      RAG grounding
      Vision and video
      Audio and speech
    Tech stack
      Python
      Jupyter Notebook
      Vertex AI
      Google Cloud
    Use cases
      Learn AI on GCP
      Build AI features
      Sample applications
    Audience
      Cloud developers
      ML practitioners

mindmap root((Google Gen AI)) Topics Gemini models RAG grounding Vision and video Audio and speech Tech stack Python Jupyter Notebook Vertex AI Google Cloud Use cases Learn AI on GCP Build AI features Sample applications Audience Cloud developers ML practitioners

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Follow step-by-step Gemini notebooks to add AI chat, summarization, or function calling to a Google Cloud app

USE CASE 2

Build a custom document search engine using the Agent Search sample notebooks as a starting template

USE CASE 3

Learn retrieval-augmented generation by running the RAG notebooks against your own documents on Vertex AI

USE CASE 4

Generate images or videos with Imagen and Veo using the vision sample notebooks

Tech stack

PythonJupyter NotebookVertex AIGoogle Cloud SDK

Getting it running

Difficulty · moderate Time to first run · 30min

Requires a Google Cloud account with billing enabled and the Vertex AI API activated.

In plain English

This repository is a collection of sample code, Jupyter notebooks, and small demo apps maintained by Google Cloud Platform to show developers how to use generative AI on Google's cloud. A Jupyter notebook is an interactive document that mixes code, written explanations, and the results of running that code in one place, which makes it a popular way to learn new techniques step by step. The material is split into folders by topic. The gemini folder holds starter notebooks and use cases for Google's Gemini family of models, including function calling and sample applications. The search folder covers Agent Search, a Google-managed solution for quickly building search engines over websites or enterprise data. The rag-grounding folder gathers examples about retrieval augmented generation and grounding, which are techniques for letting a model pull in outside information so its answers stay tied to real data. There are also folders for vision (using Imagen and Veo for image and video work), audio (using Chirp, Google's universal speech model), and setup instructions covering the Google Cloud environment, the Gen AI Python SDK, and notebook hosting on Google Colab and Workbench. The audience is developers and teams who want to build on Google Cloud's generative AI offering, particularly the Gemini Enterprise Agent Platform, described as the latest evolution of Vertex AI. The README also points to many related repositories, including the Agent Development Kit samples, the Agent Starter Pack of production-ready templates, and the Gemini Cookbook. The full README is longer than what was provided.

Copy-paste prompts

Prompt 1

Using the Gemini function calling notebooks in this repo, show me how to set up the Gen AI Python SDK, define a function schema, and call Gemini with a tool that looks up live weather data.

Prompt 2

Using the RAG grounding examples, help me build a pipeline that answers questions about my internal PDF documents by retrieving relevant chunks and passing them to Gemini on Vertex AI.

Prompt 3

Walk me through the Agent Search setup notebook so I can create a search engine over my company's internal website using Google Cloud's managed search solution.

Prompt 4

Using the Gemini vision sample notebooks, write a Python script that takes a product image and generates a marketing description using the Gemini model on Vertex AI.

Open on GitHub → Explain another repo

← googlecloudplatform on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.