Connecting to Google VertexAI
Vertex AI connections enable Flow actions to interact with Google Vertex AI resources, such as Large Language Models (LLMs).
How to create a new VertexAI Connection
When adding a Google VertexAI action (such as Agent Model or Chat Completion), you can select an existing VertexAI connection or create a new one:
- In the Flowchart, select the action you want to connect.
- In the property panel, select Connection.
- Toggle Create New Connection on.
- Fill in the required fields:
- Name: Enter a unique, descriptive name.
- Location: Enter the region name (see Locations).
- Json key: Paste the JSON key associated with your VertexAI account.
Note: Not all services are available in all locations.
Properties
| Name | Description |
|---|---|
| Name | Name of the connection. |
| Location | Region (datacenter) where your model runs. See Locations. Example: europe-west4. |
| Json key | Service Account key for authentication (paste the full content of the JSON file acquired when creating a new key). |
How to obtain the Json key
1. Account Setup
- Sign in or sign up at the Google Cloud Console with your Google account.
- Click the Project selector (top bar) → New Project → enter a name → Create.
- Ensure Billing is enabled for your project (left menu: Billing).
2. Enable VertexAI API
- In the Console left menu: APIs & Services → Library.
- Search for "Vertex AI API" (
aiplatform.googleapis.com) → click Enable.
3. Create Service Account
- In the Console left menu: IAM & Admin → Service Accounts → Create Service Account.
- Enter a name/ID and description → Create.
- In the Role field, select Vertex AI User.
- Click Continue → Done.
4. Create and Download the Json key
- In the Console left menu: Service Accounts.
- On your Service Account row, click the ⋮ button to the right → Manage keys.
- Click Add key → Create new key → select JSON → Create.
- Download the JSON key file. Paste its full content of this file into the Json key field.