This product or feature is subject to the "Pre-GA Offerings Terms" in the General Service Terms section of theService Specific Terms. Pre-GA products and features are available "as is" and might have limited support. For more information, see thelaunch stage descriptions.

Create and configure an MCP connector

The page describes how to create and configure an MCP connector in Integration Connectors. For more information about the MCP connector, seeMCP connector overview.

Before you begin

Before using the MCP connector, complete the following tasks:

• In your Google Cloud project:
  • Ensure that network connectivity is set up. For information about network patterns, seeNetwork connectivity.
  • Grant theroles/connectors.admin IAM role to the user configuring the connector.
  • Grant the following IAM roles to the service account that you want to use for the connector:
    • roles/secretmanager.viewer
    • roles/secretmanager.secretAccessor

    A service account is a special type of Google account intended to represent a non-human user that needs to authenticate and be authorized to access data in Google APIs. If you don't have a service account, you must create a service account. The connector and the service account must belong to the same project. For more information, seeCreating a service account.

  • Enable the following services:
    • secretmanager.googleapis.com (Secret Manager API)
    • connectors.googleapis.com (Connectors API)

    To understand how to enable services, seeEnabling services. If these services or permissions have not been enabled for your project previously, you are prompted to enable them when configuring the connector.

Configure the connector

A connection is specific to a data source. Itmeans that if you have many data sources, you must create a separate connectionfor each data source. To create a connection, do the following:

1. In the Google Cloud console, go to theIntegration Connectors>Connections page.

   Go to Connections

2. Select orcreate a project to use for the connector.
3. Clickadd_boxCreate Connection.

   TheCreate Connection page appears.

4. In theLocation section:
   1. ClickRegion and then select a location for the connection.

      For the list of all the supported regions, seeLocations.

   2. ClickNext.
5. In theConnection Details section, provide the following details:
   1. Connector: SelectMCP (Preview).
   2. Connector version: Select the version of the connector to use.
   3. Connection Name: Enter a name for the connection.Important: The connection name must be unique within the selected region. The connection name must adhere to the following rules:
      • The name must contain lower-case letters, numbers, or hyphens.
      • The name must begin with a letter and end with a letter or number.
      • The name must not exceed 49 characters.
   4. Description: (Optional) Enter a description for the connection.
   5. Optionally, enableCloud logging, and then select a log level. By default, the log level is set toError.
   6. Service Account: Select the service account to use for the connection. Ensure that the service account has the relevantIAM roles and permissions.
   7. MCP Endpoint: (Optional) Leave empty to use the default endpoint -/mcp. If the endpoint is same as the host address, then set this to/. Only applicable for the Streamable HTTP Server.
   8. Connection node settings: (Optional) Configure the connection node settings.
      • Minimum number of nodes: Enter the minimum number of connection nodes.
      • Maximum number of nodes: Enter the maximum number of connection nodes.

      A node is a unit (or replica) of a connection that processes transactions. More nodes process more transactions for a connection, and fewer nodes process fewer transactions. To understand how nodes affect connector pricing, see Pricing for connection nodes. If you don't enter any values, by default, the minimum number of nodes is 2 (for better availability), and the maximum number of nodes is 50.

      Note: You can customize the connection node values only if you are a Pay-as-you-go customer.
   9. ClickNext.
6. In theDestinations section, enter details of the remote host (backend system) to connect to.
   1. Destination Type: Select the destination type.
      • To specify the destination hostname or IP address, selectHost address and enter the address in theHost 1 field.
      • To establish a private connection, selectHost Address and add the endpoint attachment created for the SAP Gateway using the HTTPS protocol.
      Note:To understand how to create an endpoint attachment, seePSC service attachment andendpoint attachment. After you have created the endpoint attachment, it will be visible in theEndpoint Attachment list.

      If you want to establish a public connection to your backend systems with additional security, you can considerconfiguring static outbound IP addresses for your connections, and then configure your firewall rules to allowlist only the specific static IP addresses.

      Examples of host addresses to use:

      • https://your-service.example.com/sse (No Authentication)
      • https://your-oauth-provider.example.com/sse (OAuth 2.0 - Authorization code)
      • https://your-api-key-service.example.com/sse (API Key Authentication)
   2. To enter additional destinations, click+Add Destination.
   3. ClickNext.
7. In theAuthentication section, select and configure the authentication for the connection:

   No Authentication

   This type is used for hosts without specific authentication requirements.

   OAuth 2.0 - Authorization code

   This type is used for hosts that support the OAuth 2.0 Authorization Code grant type.

   1. Client ID:Enter the Client ID provided by your external application.
   2. Scopes:Enter the required permission scopes.
   3. Client secret:Choose theSecret Manager secret containing the client secret.
   4. Secret version:Select the Secret Manager secret version for the client secret.
   5. Enable PKCE: Optional: Enable PKCE for the auth code flow. Enable only if server supports PKCE.
   6. Authorization URL:Enter the URL for authorization.
   7. Access Token URL:Enter the URL to obtain the access token.

   For theAuthorization code authentication, after creating the connection, you must complete a few additional steps to complete the authentication configuration. For more information, seeAdditional steps after connection creation.

   Note: If you are usingAuthorization code authentication and have made any configuration changes in your application, you must re-authorize your connection. To re-authorize, seeRe-authorization for Authorization code.

   Additional steps after connection creation

   If you selected OAuth 2.0 - Authorization code for authentication, you must complete the following additional steps after creating the connection:

   1. In theConnections page, locate the newly created connection.
   2. The status for the new connection isAuthorization required.
   3. ClickAuthorization required.

      TheEdit authorization pane appears.

   4. Copy theRedirect URI value to your external application.
   5. Verify the authorization details.
   6. ClickAuthorize.

      If the authorization is successful, the connection status is set toActive in the Connections page.

   API Key Authentication

   This type is used for hosts that require an API key for authentication.

   • API key:Select the Secret Manager secret containing the API key. For example,projects/PROJECT_ID/secrets/your-api-key-secret/versions/1.
   • API key parameter name:The name of the parameter for the API key. For example,Authorization.
   • API key location:The location where the API key is sent. For example,Header.
8. ClickNext.
9. In theReview section, review the connection and authentication details. To make changes, click any of the previous sections.
10. ClickCreate.

    Integration Connectors starts creating the connection with the provided configuration. The connection creation takes a few minutes to complete. After the connection is created, the system redirects you to theConnections page.

Use the MCP connection in an integration

After you create the connection, it becomes available in Application Integration. You can use the connection in an integration through the Connectors task.

• To understand how to create and use the Connectors task in Application Integration, seeConnectors task.

Create an integration using the MCP connector

To create an integration using the MCP connector, perform the following steps:

1. Add anAPI Trigger to your integration:
   1. In the Google Cloud console, go to the Application Integration page.

      Go to Application Integration

   2. In the navigation menu, clickIntegrations. TheIntegrations List page appears.
   3. Select an existing integration or clickCreate integration to create a new one.
   4. If you are creating a new integration:
      1. Enter an integration name and description in theCreate Integration pane.
      2. Select a region for the integration.TheRegions dropdown only lists the regions provisioned in your Google Cloud project. To provision a new region, ClickEnable Region. SeeEnable new region for more information.
      3. Optional. Select a service account for the integration.The option to select a service account is displayed only if you have enabled integration governance for the selected region.
      4. ClickCreate. The newly created integration opens in the integration editor.
   5. In the integration editor navigation bar, clickTriggers to view the list of available triggers.
   6. Click and place theAPI trigger element to the designer.
2. AddData Mapping if required.
3. Go toTasks and choose one of the connectors listed above orthe one you created, if applicable.
4. ClickConfigure connector and choose actions. The default settings are pre-selected.

   A dropdown will show various actions, which are tools in this context.

   

   1. Configure the required input parameters for the action. You can set these as integration inputs, use variables from previous tasks, or specify default values directly.
   2. Pass the required arguments and test the connection.

      

For the list of supported entities and operations, seeSupported entities and operations. For information about the supported connector-specific actions, seeSupported actions.

Use case example in Application Integration

The following examples demonstrate how to use the Connectors task in Application Integration to perform operations using your configured MCP connection.

Example - Jira: Create Issue

To create an issue in Jira using a MCP connection:

1. In theConfigure connector task dialog, select your Jira MCP connection (e.g.,jira-connection).
2. ClickActions.
3. Select thecreate_issue action.
4. ClickDone.
5. In theTask Input section, pass the following arguments:

   {"fields":{"project":{"key":"YOUR_PROJECT_KEY"},"summary":"New issue created from Integration Connectors","issuetype":{"name":"Task"},"description":"This is a test issue created via the MCP Jira connector."}}

Create connections using Terraform

You can use theTerraformresource to create a new connection.

To learn how to apply or remove a Terraform configuration, seeBasic Terraform commands.

To view a sample terraform template for connection creation, seesample template.

When creating this connection by using Terraform, you must set the following variables in your Terraform configuration file: