Claude Google Cloud BigQuery connectorWhat Claude can do in your Google Cloud BigQuery account.
The Claude Google Cloud BigQuery connector exposes 5 tools. 4 read the structure of your projects, 1 runs any SQL, including statements that change data. Below: what Claude can explore, what that one query tool can do, and the roles and limits Google sets.
Verified Trustpilot reviews · AI, automation & growth agency
What changes for the people who wait on the data team
Most people who need a number from BigQuery do not write SQL. With the connector, you ask Claude in plain words, it finds the right dataset and table, reads how they are built, then writes and runs the query itself. The answer comes back in the chat, next to the question you asked.
Finding where the data lives. A new analyst has no idea which project holds the orders. list_dataset_ids and list_table_ids walk the project and its datasets, so Claude can point at the right table instead of guessing.
Checking a table before trusting it. get_dataset_info and get_table_info return the metadata of a dataset or a table, which is how Claude learns how the data is built before writing anything.
Getting the figure, or a forecast. execute_sql runs the query and returns the result. Google also presents BigQuery's advanced capabilities, forecasting among them, as a way to get higher-order insights.
What Claude will not do here. The query tool runs any statement BigQuery accepts, so it can insert, update, delete or create just as easily as it can read: it is the only one Google does not label read-only. It cannot reach external tables stored in Google Drive, and a query stops after three minutes. Nothing runs on a schedule either: no tool fires when a table is refreshed. For recurring loads, the Google Cloud BigQuery Make integration covers a different job, background scenarios rather than conversation.
Five words before you plug it in
The vocabulary you will meet while connecting BigQuery.
- Connector
- The link you set up once between Claude and an account you already have, so Claude can work in that account while it answers you.
- Tool
- One named thing the connector lets Claude do. Claude picks the ones it needs by itself, and the directory sheet lists them by name.
- Authorization
- The service's own sign-in screen, where you hand Claude the access it will use later. Given once per person, and taken back the same way.
- Approval
- The confirmation Claude waits for before it goes through with a change to your account, shown in the chat at the moment it matters.
- MCP
- The shared standard connectors are built on: it is what lets an assistant such as Claude talk to an outside service like BigQuery.
Connecting BigQuery to Claude in three steps
- 01
Find BigQuery in Claude
In Claude's settings, open Customize, then Connectors, and look for Google Cloud BigQuery. On a Team or Enterprise workspace, an Owner or Primary Owner has to switch the connector on before each member can sign in.
- 02
Start the connection
Click Connect on its row, then sign in with your Google account in the window Google opens itself. If the link breaks later, Disconnect and connect again.
- 03
Read the authorization screen
Read Google's screen before you accept. It belongs to the service, not to Claude, and it decides what the access covers. You can also withdraw that access later from your Google account.
The 5 tools, sorted by what they do
Google Cloud BigQuery gives Claude 5 tools: 4 that read your account, 1 that change something in it.
Two groups: what Claude reads about your projects, and the one tool that can also change them. Tool names stay as Claude shows them.
- 4 read
- 1 write
What Claude reads (4)
4 toolsFour tools that describe your projects, datasets and tables without touching the data.
list_dataset_ids
Returns the dataset IDs, and the BigLake namespaces, of a Google Cloud project. It is the map Claude reads first in a project it does not know yet.
get_dataset_info
Brings back the metadata of one dataset or BigLake namespace, so Claude knows what it is dealing with before opening its tables.
list_table_ids
Lists the table IDs inside a dataset or BigLake namespace. Claude uses it to narrow a vague question down to the table that holds the answer.
get_table_info
Returns the metadata of a BigQuery or BigLake table. That is how Claude learns what the table is before it drafts a query against it.
What Claude can change (1)
1 toolOne tool that runs any SQL BigQuery supports. No source describes a confirmation specific to it: the general rule below applies.
execute_sql
Approval: see the ruleRuns a SQL query in the project and hands back the result. Google's own reference lists inserts, updates, deletes and creation statements among what it accepts, and advises the read-only variant whenever possible.
What Claude checks with you before acting
By default, Claude stops and asks for your go-ahead before each action it takes on an account on your behalf. The request shows up in the chat, right when it matters.
On Team and Enterprise, the owners decide whether a member may let some actions through without being asked every time. They can also limit what a connector may do for the whole organisation, keeping reading open and closing writing, for instance. Google adds its own lock: IAM allow and deny policies, and a deny policy aimed at read-write tools shuts the query tool. And Claude only has your rights: what your Google identity cannot reach, it cannot reach either.
Which Claude plans include it
Out of the 819 sheets in the official directory, none shows availability by plan. The connector-by-connector answer is not published anywhere: it is a real gap in the catalogue.
The general rule is published: remote connectors are open to every user on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner turns the connector on for the organisation before members can connect. For the current state of this one, the connector's sheet in the official directory is the place to check. BigQuery's own quotas are Google's business and stay on the Google side.
Where this connector stops
A connector is not an automation. Claude calls these tools while it answers you: nothing starts when a table is loaded, a job finishes or a number crosses a threshold.
Everything here comes from Google's own documentation: neither Claude's help centre nor Anthropic's developer docs have a page on this connector. The tool list is an observed floor, not a promise, since an admin can open actions no public sheet shows. The partner badge is not a security audit, and Anthropic writes on every sheet that it does not choose a publisher's tools nor guarantee how they behave. Google also documents a read-only query tool that is not on the directory sheet, so here every query goes through the one that can change data. More connectors are listed on the Integrations hub.
Need help connecting Google Cloud BigQuery to Claude?
A person reads every message.