Skip to content

Quick Start

Jump to bottom
jorsm edited this page May 29, 2026 · 3 revisions

Quick Start

Get started quickly with Vertex AI models in VS Code.

Table of Contents

1. Install the extension

Install Vertex AI Models Chat Provider from the VS Code Marketplace.

2. Authenticate with Google Cloud

This extension supports two authentication methods:

Option A: Standard Credentials (ADC)

Uses your local gcloud identity. Run:

gcloud auth application-default login

Option B: Service Accounts (Recommended for Teams)

Use a JSON key without installing the gcloud SDK:

  1. Run command: Vertex AI: Paste Service Account JSON Key
  2. Paste your JSON and give it a name.

See Service Account Authentication for details.

3. Enable Vertex AI

In the Google Cloud Console, make sure the following are configured for your project:

  • Vertex AI API enabled (aiplatform.googleapis.com)
  • Your account has the Vertex AI User role (roles/aiplatform.user)
  • Anthropic Claude models are enabled via the Vertex AI Model Garden if you plan to use them

4. Set your GCP Project ID in VS Code

Open VS Code Settings and set:

{
  "vertexAiChat.projectId": "YOUR_PROJECT_ID"
}

Use a workspace-level setting in .vscode/settings.json if you want different workspaces to use different projects.

5. Open the Chat panel

Open the VS Code Chat panel and select the Google Cloud Vertex AI provider.

Once your project is configured, the extension automatically discovers available models in your Google Cloud project and region.

6. Refresh models manually (optional)

If you update model access or project permissions, run the command:

  • Vertex AI Models Chat Provider: Refresh Models

7. Open usage and billing dashboard

Open the built-in dashboard by clicking the cost status bar item in VS Code.

This gives you local cost estimates and token usage details for your current session.

Clone this wiki locally