Skip to main content
Connections are added inside an Access bundle. At claude.ai/admin-settings/claude-tag, open Access bundles in the left navigation, click into a bundle (or Create one), and go to its Credentials tab.
Connecting BigQuery lets Claude run queries against your datasets from any channel under the bundle’s scope. Add it as a custom credential with Custom tool; BigQuery has no preset button in the picker. This is an HTTP API connection, not a personal claude.ai connector. Pair it with a plugin that covers BigQuery so Claude knows how to form and run queries; without one, Claude can reach the API but has to work out the request shape on its own. See Attach plugins.

​
Create the credential in Google Cloud

Create a dedicated service account for the agent in the Google Cloud project that holds your BigQuery data, then create a JSON key for it. Google’s guides cover creating a service account and creating a service account key.

​
Grant access to specific datasets

You scope what Claude can read on the Google Cloud side, through the service account’s role grants. The connection itself has no dataset setting. Grant the service account two roles:
  • BigQuery Data Viewer (roles/bigquery.dataViewer) on each dataset Claude should query. Grant it on the specific datasets, not on the project, so Claude can read only those datasets.
  • BigQuery Job User (roles/bigquery.jobUser) on the project, so the service account can run query jobs.
Together the two grants let Claude run read-only queries against those datasets. To widen or narrow access later, edit the dataset grants in Google Cloud; the connection needs no change.

​
Add the connection to a bundle

In the bundle, click Connect next to Custom tool and choose GCP access token (with Service Account Key). Agent Proxy exchanges the service-account key for an access token and injects it at the network boundary; the model and the sandbox are not given the key. See how Agent Proxy works.

​
Verify the connection

In a channel under the bundle’s scope, in a new thread:
BigQuery appears in the list once the connection is live. New threads pick up the connection on their own; in an existing thread, ask Claude to use the service by name. Then confirm a query runs against a dataset you granted:

​
Allow the connection through a VPC Service Controls perimeter

If the Google Cloud project that holds your BigQuery data is inside a VPC Service Controls perimeter, Claude’s queries fail with Request is prohibited by organization's policy and a vpcServiceControlsUniqueIdentifier in the error details. To let them through, add an ingress rule that admits the service account you created for Claude. If you allowlist Anthropic’s published egress IP range in an access level, Claude’s queries still fail, because they reach your perimeter from inside Google Cloud rather than from that range. The source your perimeter sees is a Google Cloud project that Anthropic owns and can change without notice, so don’t admit that project by its number. Add an ingress rule to the perimeter with these settings:
  • For the identity, admit the service account you created for Claude.
  • For the source, allow any source (an access level of *).
  • For the target, allow the BigQuery API (bigquery.googleapis.com) on the project inside the perimeter.
To confirm the rule works, ask Claude to run the query that failed. The query returns results.

​
What the VPC Service Controls ingress rule allows

The ingress rule lets requests authenticated as the service account you created for Claude cross the perimeter, and only to reach BigQuery in the project the rule names. It doesn’t give the service account access to any data. The dataset roles you granted still decide which datasets Claude can read, so keep those grants narrow. Once the connection works, delete the key file that Google Cloud downloaded to your machine when you created the key. With this rule in place, the perimeter doesn’t stop a leaked key for this service account, because a request authenticated with any of the account’s valid keys passes the rule from any source. Agent Proxy holds the key you uploaded and Claude never sees it. To limit that risk further, don’t create more keys for the service account.