@core-ai/google-vertex

Google on Vertex AI provider package for @core-ai/core-ai. It uses the @google/genai client in Vertex AI mode and shares chat, embedding, image generation, structured-output, and reasoning behavior with @core-ai/google-genai.

Installation

npm install @core-ai/core-ai @core-ai/google-vertex zod

Usage

import { generate } from '@core-ai/core-ai';
import { createGoogleVertex } from '@core-ai/google-vertex';

const googleVertex = createGoogleVertex({
    projectId: process.env.GOOGLE_VERTEX_PROJECT,
    region: 'europe-west1',
});
const model = googleVertex.chatModel('gemini-2.5-flash');

const result = await generate({
    model,
    messages: [{ role: 'user', content: 'Hello!' }],
});

console.log(result.content);

Authentication

By default, the provider uses Google Application Default Credentials (ADC):

const googleVertex = createGoogleVertex({
    projectId: process.env.GOOGLE_VERTEX_PROJECT,
    region: 'europe-west1',
});

To authenticate with an explicit service account, parse its JSON key and pass it as credentials:

const rawCredentials = process.env.GOOGLE_APPLICATION_CREDENTIALS_JSON;
const credentialsJson = rawCredentials?.trim().startsWith('{')
    ? rawCredentials
    : Buffer.from(rawCredentials ?? '', 'base64').toString('utf8');

const googleVertex = createGoogleVertex({
    projectId: process.env.GOOGLE_VERTEX_PROJECT,
    region: 'europe-west1',
    credentials: JSON.parse(credentialsJson),
});

You can also inject a preconfigured GoogleGenAI client:

import { GoogleGenAI } from '@google/genai';

const googleVertex = createGoogleVertex({
    client: new GoogleGenAI({
        vertexai: true,
        project: 'my-project',
        location: 'europe-west1',
    }),
});

Models and regions

Pass Vertex AI model IDs to the corresponding model factory:

const chat = googleVertex.chatModel('gemini-2.5-flash');
const embeddings = googleVertex.embeddingModel('gemini-embedding-001');
const image = googleVertex.imageModel('gemini-2.5-flash-image');

Model availability varies by region. A provider instance targets one region, so create separate providers when models are hosted in different locations.

Image generation

import { generateImage } from '@core-ai/core-ai';

const result = await generateImage({
    model: googleVertex.imageModel('gemini-2.5-flash-image'),
    prompt: 'A watercolor robot in a mountain cabin at sunrise',
    size: '1024x1024',
});

console.log(result.images[0]?.base64);

Gemini image models use native multimodal generation and return base64 image data.

Provider options and capabilities

This provider shares its model behavior with @core-ai/google-genai. Provider options remain namespaced under google. @core-ai/google-vertex re-exports the applicable provider-option schemas, capability helpers, and reasoning metadata types.