Skip to main content
Back to the BigQuery MCP server

MCP Gateway · Claude Cowork

BigQuery icon

Connect BigQuery to Claude Cowork

Some warehouse work is a whole afternoon: build the weekly KPI summary, reconcile two sources, document a dataset nobody wrote up. In Cowork, Claude can carry each of those through many queries as one task, through the Speakeasy MCP Gateway. Agent hooks send each prompt to Speakeasy before the step runs, and OTLP export ties each session to a person.

Category
Data & analytics
Tools
6
Authentication
OAuth
Claude Cowork
Cowork tab in Claude Desktop

Behind the MCP Gateway

Run BigQuery MCP behind the MCP Gateway

One URL for every agent, sign-in through your IdP, RBAC down to the tool, and every BigQuery call inspected and logged.

Deploy on MCP Gateway

Before you start

What BigQuery needs first

A Google Cloud project administrator with IAM admin access does the BigQuery side of the setup.

  • You need a Google Account that can administer the project and create OAuth credentials.
  • Granting roles needs IAM admin access, such as Project IAM Admin.
  • Billing is optional for initial setup.
  • External apps in Testing allow at most 100 test users, and authorizations expire after seven days.

In BigQuery

  1. Enable the BigQuery API in the project.
  2. Grant each user MCP Tool User, BigQuery Job User, and BigQuery Data Viewer.
  3. Configure the OAuth consent screen as Internal or External.
  4. For External apps, add the BigQuery scope and test users, or publish the app.
  5. Create an OAuth client and copy its credentials.
Full BigQuery setup guide in the docs

Set up

Add BigQuery to Claude Cowork in 4 steps

  1. 01

    Add BigQuery from the catalog

    An admin finds BigQuery in the MCP Catalog (MCP Gateway > MCP > Catalog) and clicks Add. Speakeasy registers the vendor's server as a remote MCP server and proxies sessions to it. Nothing is copied or re-hosted.

    BigQuery uses OAuth. Servers that support dynamic client registration get their OAuth configuration applied automatically.

    Where MCP-scoped policies are enabled, the install offers a recommended risk policy for the new server that detects secrets and PII and inspects tool requests and responses.

  2. 02

    Choose who can use BigQuery

    Give teams the mcp:connect scope on the BigQuery server through team access or a role. Claude Cowork connects to the BigQuery server's own URL, and the server's authorization decides who can connect and which tools they can call.

  3. 03

    Connect Claude Cowork

    Sync the organization's plugin marketplace into Claude, then make the plugins available in Cowork.

    New MCP servers join the project's default plugin automatically. Each server's overview page shows which plugin it is published to, with Publish or Update to change it.

    Add the importing admin's GitHub username as a marketplace collaborator and accept the invite. Only that admin needs it.

    In Claude Organization settings > Plugins & skills, select Add > Sync from GitHub and pick the organization's plugin repository. The Claude GitHub App must be installed on it.

    Mark the observability plugin Required, and set the plugin that carries this server to Available to install or Installed by default.

    In the Cowork tab's Monitoring settings, set the OTLP endpoint, protocol and headers.

  4. 04

    Sign in and use BigQuery from Claude Cowork

    The first time Claude Cowork connects, it opens a browser so the person can sign in to Speakeasy through your identity provider.

    BigQuery needs its own OAuth grant, so each person connects it once. Until they do, BigQuery calls fail while the rest of the gateway keeps working.

    From then on, BigQuery's tools appear in Claude Cowork, limited to what the person's access allows, and every call lands in the gateway's tool logs.

Cowork reads the synced repository, so a new session picks up plugin changes. Members who receive plugins through Cowork don't need a GitHub invitation.

Use cases

What Claude Cowork users do with BigQuery

Build the weekly KPI summary

Cowork finds the source tables, checks their columns, runs a read-only query per metric and writes the results up as a summary for the team.

list_table_idsget_table_infoexecute_sql_readonly

Reconcile two sources of truth

When billing and product data disagree, Cowork queries both tables, matches records and lists the rows that differ along with a likely cause.

execute_sql_readonlyget_table_info

Write a data dictionary

Cowork reads a dataset's metadata and every table in it, then drafts a plain-language dictionary of tables and fields for new analysts.

get_dataset_infolist_table_idsget_table_info

Tools

6 tools

Each BigQuery tool can be allowed or denied per team and role at the MCP Gateway before Claude Cowork sees it.

list_dataset_ids

List BigQuery dataset IDs in a Google Cloud project.

get_dataset_info

Get metadata information about a BigQuery dataset.

list_table_ids

List table ids in a BigQuery dataset.

get_table_info

Get metadata information about a BigQuery table.

execute_sql_readonly

Run a read-only SQL query in the project and return the result. Prefer this tool over execute_sql if possible.

execute_sql

Run a SQL query in the project and return the result. Prefer the execute_sql_readonly tool if possible.

Control

What Speakeasy can see and stop in Claude Cowork

The gateway covers every client the same way. The rest depends on what Claude Cowork exposes to Speakeasy.

Tool logs

Gateway

Every BigQuery call is logged with the person and the client that made it.

Access by role

Gateway

People reach only the BigQuery tools their team or role allows.

Request and response blocking

Gateway

A risk policy scoped to BigQuery can reject a tool request before it reaches the server, or withhold a result from Claude Cowork. MCP-scoped policies are enabled per organization.

Prompt blocking

Claude Cowork

Agent hooks send each prompt to Speakeasy, which allows or denies it before the step runs.

Session telemetry

Claude Cowork

Cowork's OTLP export ties sessions to people and feeds observability.

Anthropic inference hooks

Claude Cowork

On Claude Enterprise, prompts can be checked before inference (beta).

  • Neither the device agent nor MDM reaches Cowork, so the Claude-side setup always applies.
  • Personal Claude accounts can't be governed in Cowork.

Questions

Frequently asked questions

Can I keep Cowork read-only in BigQuery?
Yes. Allow execute_sql_readonly and deny execute_sql for the roles that use Cowork, so the queries it runs can read data but not change it.
Why does BigQuery ask Cowork users to sign in again each week?
If the Google OAuth app is External and in Testing, each user's authorization expires after seven days. Ask your Google Cloud admin about publishing the app.
How do Claude Cowork users get BigQuery?
An admin adds BigQuery from the Speakeasy MCP Catalog and grants access by team or role. Sync the organization's plugin marketplace into Claude, then make the plugins available in Cowork. Each person then sees only the BigQuery tools their access allows.
Are BigQuery calls from Claude Cowork logged?
Yes. BigQuery calls go through the MCP Gateway, which logs each tool call and attributes it to the person and the client that made it.

Connect more

BigQuery in other clients

AI everywhere.

Control here.