Skip to content

Latest commit

 

History

History
40 lines (27 loc) · 1.74 KB

README.md

File metadata and controls

40 lines (27 loc) · 1.74 KB

Using the Google Cloud Vertex AI PaLM API to create kid stories

This application takes advantage of the PaLM API, in Google Cloud's Vertex AI, to generate kid stories. Users select or invent a character, a setting (location, time), and a general plot. This information is sent to the PaLM API, via a prompt, that asks for a structured story to be generated. The frontend part is using vanilla JavaScript, with the Shoelace web components for the form inputs. And the backend is developed with the Micronaut framework, the Apache Groovy programming language, and is deployed on Cloud Run.

Bed-time stories

Note: As of this writing, there is no Java client library to access the PaLM API. So this application is calling the REST endpoint directly. However, to authenticate to that application, I had to use the Google Auth client library, as explained in this article.

Building and deploying

You'll need to have the Vertex AI API enabled, as well as the various APIs needed by Cloud Run, Cloud Build, and Artifact Registry.

Building with Cloud Build

gcloud builds submit -t us-central1-docker.pkg.dev/genai-java-demos/containers/bedtimestories:v1

Note: change the genai-java-demos with your own project ID.

Deploying to Cloud Run

gcloud run deploy bedtimestories --image=us-central1-docker.pkg.dev/genai-java-demos/containers/bedtimestories:v1

Note: change the genai-java-demos with your own project ID.


This is not an official Google product.