Google Cloud Services - Gen AI

This extension allows to inject a com.google.genai.Client object, from the Google Gen AI SDK, inside your Quarkus application. The client can talk to Vertex AI (default) or to the Gemini Developer API.

It replaces the Vertex AI extension, which relies on Vertex AI SDK generative APIs that are deprecated by Google.

Be sure to have read the Google Cloud Services extension pack global documentation before this one, it contains general configuration and information.

Bootstrapping the project

First, we need a new project.Create a new project with the following command (replace the version placeholder with the correct one):

mvn io.quarkus:quarkus-maven-plugin:<quarkusVersion>:create \
    -DprojectGroupId=org.acme \
    -DprojectArtifactId=genai-quickstart \
    -Dextensions="resteasy-reactive-jackson,quarkus-google-cloud-genai"
cd genai-quickstart

This command generates a Maven project, importing the Google Cloud Gen AI extension.

If you already have your Quarkus project configured, you can add the quarkus-google-cloud-genai extension to your project by running the following command in your project base directory:

./mvnw quarkus:add-extension -Dextensions="quarkus-google-cloud-genai"

This will add the following to your pom.xml:

<dependency>
    <groupId>io.quarkiverse.googlecloudservices</groupId>
    <artifactId>quarkus-google-cloud-genai</artifactId>
</dependency>

Choosing the backend

By default, the client uses Vertex AI with the Google Cloud credentials and project ID configured as described in the global documentation. You can set the region with:

quarkus.google.cloud.genai.location=us-central1

If no project ID or location is configured, the SDK falls back to the GOOGLE_CLOUD_PROJECT and GOOGLE_CLOUD_LOCATION environment variables.

To use the Gemini Developer API instead, configure an API key, no Google Cloud credentials are needed:

quarkus.google.cloud.genai.api-key=<your-api-key>

The backend is selected as follows: quarkus.google.cloud.genai.vertex-ai if set, otherwise the Gemini Developer API when an API key is configured, otherwise Vertex AI. Setting an API key together with vertex-ai=true fails at startup. The API key is a secret: prefer an environment variable or a secret source over a plain application.properties.

Some example

This is an example usage of the extension: we create a REST resource with a single endpoint that takes a prompt as query parameter and generates a response using Gemini Flash.

import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.QueryParam;

import com.google.genai.Client;

@Path("/genai")
public class GenAIResource {
    @Inject
    Client client;

    @GET
    public String generate(@QueryParam("prompt") String prompt) {
        return client.models.generateContent("gemini-2.5-flash", prompt, null).text();
    }
}

Migrating from the Vertex AI extension

Replace the quarkus-google-cloud-vertex-ai dependency by quarkus-google-cloud-genai and the quarkus.google.cloud.vertexai. properties by quarkus.google.cloud.genai.. Then replace the VertexAI and GenerativeModel usages by the Client API, see the Google migration guide.

Configuration Reference

Configuration property fixed at build time - All other configuration properties are overridable at runtime

Configuration property	Type	Default
`quarkus.google.cloud.genai.api-key` Gemini Developer API key. When set, the client uses the Gemini Developer API instead of Vertex AI, unless `vertex-ai` is set to `true`. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_API_KEY`	string
`quarkus.google.cloud.genai.vertex-ai` Force the backend: `true` for Vertex AI, `false` for the Gemini Developer API. Defaults to Vertex AI unless an API key is configured. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_VERTEX_AI`	boolean
`quarkus.google.cloud.genai.location` Google Cloud region used by Vertex AI, for example `global` or `us-central1`. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_LOCATION`	string
`quarkus.google.cloud.genai.base-url` Override the base URL of the API. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_BASE_URL`	string
`quarkus.google.cloud.genai.api-version` Override the API version. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_API_VERSION`	string
`quarkus.google.cloud.genai.timeout` Timeout of HTTP requests. Environment variable: `QUARKUS_GOOGLE_CLOUD_GENAI_TIMEOUT`	Duration

Configuration property

Type

Default

quarkus.google.cloud.genai.api-key

Gemini Developer API key. When set, the client uses the Gemini Developer API instead of Vertex AI, unless vertex-ai is set to true.

Environment variable: QUARKUS_GOOGLE_CLOUD_GENAI_API_KEY

string

quarkus.google.cloud.genai.vertex-ai

Force the backend: true for Vertex AI, false for the Gemini Developer API. Defaults to Vertex AI unless an API key is configured.

Environment variable: QUARKUS_GOOGLE_CLOUD_GENAI_VERTEX_AI

boolean

quarkus.google.cloud.genai.location