Integration · Company data

Connect Claude & ChatGPT to Google BigQuery

Ask your AI a question about your data and let it write and run the SQL in BigQuery. It can list your projects, datasets and tables, read a table's schema before it queries, preview rows and follow up on long-running jobs. It can also create datasets and tables, stream in rows and update descriptions and labels, if you switch those tools on.

Connect Google BigQuery free

Free plan, no credit card. Takes about a minute.

Try asking

“Which tables are in the sales dataset, and what columns does orders have?”
The AI lists the tables and views in the dataset, then reads the orders table's schema, row count and size.
“How much revenue did we book per month this year?”
It checks the table schema, writes a GoogleSQL query with a monthly total and runs it in your project.
“How much data would that query scan before I run it?”
It runs the query as a dry run, which estimates the bytes scanned without running it.

What your AI can do

Queries and results

Run SQL queries with a dry run or a cap on bytes billed, fetch results page by page and read or cancel jobs.

Datasets and tables

Browse projects, datasets, tables, views and saved routines, read schemas and preview rows.

Create and update

Create datasets and tables, stream in rows, update descriptions, labels and columns, and delete tables.

“Show me the first 20 rows of the events table.”
It reads rows straight from the table without running a query.
“Did last night's query jobs fail?”
It lists recent jobs in the project filtered by state and reads the status and errors of the one you pick.
“Create a table called leads in the marketing dataset with name, email and created_at.”
It creates an empty table with that column schema. Rows can then be streamed in with insert_rows.

18 tools for Google BigQuery

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

cancel_jobDeletesAsk BigQuery to cancel a running job by `projects_id` and `jobs_id`. Work it already did is not undone.
create_datasetChanges dataCreate a new empty dataset in project `projects_id`. Set the dataset ID in `dataset_reference`, plus `location` and `description`.
create_tableChanges dataCreate a new empty table in a dataset by `projects_id` and `datasets_id`. Put the table ID in `body.tableReference` and the columns in `body.schema`.
delete_tableDeletesPermanently delete a table and all its data by `projects_id`, `datasets_id` and `tables_id`. This cannot be undone once the time-travel window passes.
get_datasetGet one dataset's details by `projects_id` and `datasets_id`, including location, description, labels and access settings.
get_jobGet one job's status, statistics and any errors by `projects_id` and `jobs_id`.
get_query_resultsGet the results of a query job by `projects_id` and `queries_id` (the job ID from run_query), page by page. Use when run_query returned before the query finished.
get_routineGet one routine's definition and arguments by `projects_id`, `datasets_id` and `routines_id`.
get_tableGet one table's details by `projects_id`, `datasets_id` and `tables_id`, including its column schema, row count and size. Check this before writing a query.
insert_rowsChanges dataStream new rows into a table by `projects_id`, `datasets_id` and `tables_id`; pass `rows` as a list of objects with a `json` field holding each row.
list_datasetsList the datasets in a project (`projects_id`). Use `filter` to narrow by label and `all` to include hidden datasets.
list_jobsList recent BigQuery jobs (queries, loads, exports) in project `projects_id`, filtered by `state_filter` or creation time.
list_routinesList the saved routines (user-defined functions and stored procedures) in a dataset by `projects_id` and `datasets_id`.
list_tablesList the tables and views in a dataset by `projects_id` and `datasets_id`.
preview_table_rowsRead rows straight from a table by `projects_id`, `datasets_id` and `tables_id`, optionally only `selected_fields` and from `start_index`. Free and fast for a quick look; use run_query to filter or aggregate.
run_queryChanges dataRun a SQL `query` (GoogleSQL by default) in project `projects_id` and return the results if it finishes within `timeout_ms`. Queries are billed by bytes scanned; set `dry_run` to estimate cost first, or `maximum_bytes_billed` to cap it. Statements like INSERT, UPDATE, DELETE or DROP change data permanently.
update_datasetChanges dataUpdate a dataset's description, labels, friendly name or default expiration by `projects_id` and `datasets_id`. Only the fields you pass change.
update_tableChanges dataUpdate a table's description, labels, expiration or add columns by `projects_id`, `datasets_id` and `tables_id`. Only the fields you pass in `body` change.

Set up in three steps

1
Pick the app
Create a free PipMCP account and choose Google BigQuery from the app list.
2
Paste your key
Paste the OAuth client ID and client secret from your own Google Cloud project, plus the ID of the project that holds your data, then sign in with Google once. In the Google Cloud console create or pick a project and enable the BigQuery API. Set up the OAuth consent screen, then go to Credentials > Create credentials > OAuth client ID, choose Web application and add the redirect URI https://pipmcp.com/oauth/callback.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
Click your name, then Settings › Connectors › Add custom connector. Paste your link as the Remote MCP server URL. Switch it on from the + menu in a chat.

Questions

What can the AI do in Google BigQuery?

It can list projects, datasets, tables, routines and jobs, read table schemas, preview rows and run SQL queries. It can create datasets and tables, stream in rows, update tables and datasets, cancel jobs and delete tables, if you switch those tools on.

Does the AI see my Google BigQuery credentials?

No. Your Google BigQuery credentials are encrypted at rest and never shown to the AI. After you save them, they are not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, so you can start with browsing and previews. A SQL statement like DELETE or DROP changes data permanently, and deleting a table removes it and its data, so those actions can require your approval before they run. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT go to Settings › Apps & Connectors › Advanced and turn on Developer mode, then add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

Google BigQuery connects with OAuth, and this OAuth connection needs the PipMCP Pro plan. PipMCP also has a free plan with no credit card, but it does not include this connection. Paid plans bill per completed task. You also need your own Google BigQuery account.

Will the AI's queries add to my BigQuery bill?

Queries run in your own Google Cloud project, and BigQuery bills them by bytes scanned. The AI can do a dry run to estimate the bytes first, or set a maximum bytes billed so a query stops instead of scanning more. Previewing table rows reads straight from the table without running a query.