Skip to main content

Manage exports using Funnel MCP

Create, run and turn exports on or off using Funnel MCP.

Ask an AI client connected to Funnel MCP, such as Claude, ChatGPT, Claude Code or Cursor, to set up a scheduled export to Google Sheets, Google BigQuery, Snowflake or Google Cloud Storage. You can also ask about your existing exports, turn an export off or on and run an export right away.

​

Supported exports

Funnel MCP can create exports to Google Sheets, Google BigQuery, Snowflake and Google Cloud Storage. Set up exports to other destinations in Funnel.

Destination

Create using Funnel MCP

Google Sheets

Yes

Google BigQuery

Yes

Snowflake

Yes

Google Cloud Storage

Yes

Looker Studio, Tableau and Power BI views

No. You can list and read them.

Microsoft Excel, Meridian, Microsoft Azure Blob Storage, SFTP, PostgreSQL, Amazon S3 and Amazon Redshift

No. Set them up in Funnel.

Using Funnel MCP, you can:

​

  • Ask about your exports and views, and query the data an export sends.

  • Create an export to Google Sheets, Google BigQuery, Snowflake or Google Cloud Storage.

  • Turn an export off or on.

  • Run an export now, for its own date range or for the dates you choose.

​

Using Funnel MCP, you can't:

​

  • Create Looker Studio, Tableau or Power BI views.

  • Edit or delete an export. Make those changes on the export in Funnel.

  • Choose a schedule other than daily.

​

Before you begin

  • You have an AI client connected to Funnel MCP. To connect one, see Quick start guide: Using Funnel MCP.

  • You have access to the workspace. This is enough to ask about exports and views.

  • To create, turn off, turn on or run an export, you have the Edit exports permission in the workspace. If you don't, Funnel MCP tells you. Ask a subscription owner for access. For more, see Workspace: User roles.

  • To create an export, you have an account that can write to the destination: the spreadsheet, the BigQuery project, the Snowflake database or the Cloud Storage bucket.

​

Find and inspect your exports

Ask your AI client which exports and views the workspace has, or ask about one export to see what it includes and when it runs. For example:

​

  • Which exports do we have, and which ones are turned off?

  • What does the Weekly spend export include?

  • Show me the data the Weekly spend export sends for last week.

​

The client lists exports and views with list_exports and reads one with get_export. get_export also returns a ready-made query_data request, so the client can show you the export's data in the conversation. If some of the export's settings can't be reproduced, the client tells you the result is approximate. Views have no date range, so the client asks you which dates to query.

​

Create an export

  1. In your AI client, describe the export: the dimensions and metrics, the destination, the date range and the time it should run. For example: Export cost by date and connector for the last 30 days to BigQuery, daily at 06:00.

    ​

    Note: Google BigQuery, Snowflake and Google Cloud Storage exports must include the Date dimension.

    ​

  2. If you don't have a suitable connection to the destination, the client gives you a link to sign in through Funnel. Sign in with an account that can write to the destination, then ask the client to continue. Never type passwords or other credentials in the conversation.

    ​

  3. Review the change the client shows you, including the destination, the fields, the dates and the schedule.

    ​

  4. Approve the change in the conversation within 30 minutes. If your client asks for its own approval to use the Funnel MCP tool, allow it.

​

A new export is turned on and runs once right away. To create it without the first run, ask the client not to run it now. If you don't approve the change within 30 minutes, it expires. Ask the client to prepare it again.

​

Destinations

Destination

What Funnel writes

Google Sheets

A tab in a new or existing spreadsheet. Each run replaces everything in the tab. If you don't give a spreadsheet URL, the first run creates the spreadsheet. You can't use a tab that another export already writes to.

Google BigQuery

A table in your own Google Cloud project, written in monthly parts. The default table name is funnel_data.

Snowflake

A table in an existing database in your Snowflake account, written in monthly parts. The default table name is funnel_data.

Google Cloud Storage

Gzipped CSV files in your bucket, one file per month. You can choose a folder in the bucket.

Schedule and dates

  • Schedule: The export runs daily at the time and timezone you choose. The default is 06:00 UTC.

  • Dates: Use a rolling range, such as the last 7 days or the last 3 months, or a fixed start and end date. The default is the last 30 days.

​

Turn an export off or on

  1. In your AI client, ask to turn off or turn on an export. For example: Turn off the Daily cost to Sheets export.

    ​

  2. Review the change the client shows you.

    ​

  3. Approve the change in the conversation within 30 minutes.

​

A turned-off export stops running until you turn it on again.

​

Run an export now

  1. In your AI client, ask to run an export. For example: Run the BigQuery cost export now for 1 to 15 October. Without dates, the export runs for its own date range.

    ​

  2. Review the change the client shows you.

    ​

  3. Approve the change in the conversation within 30 minutes.

​

You can't run an export that is turned off. Turn it on first.

​

Ask where you can send data

Ask your AI client Where can I send my data? The client lists the destinations with list_export_destinations. For destinations you set up in Funnel, it gives you a link to create the export in Funnel.

​

Export tools available in Funnel MCP

Tool

Type

Description

list_exports

Read

Lists the exports and views in the workspace.

get_export

Read

Returns the details of one export, including a ready-made query_data request for the same data.

list_export_destinations

Read

Lists export destinations. For destinations you set up in Funnel, it returns a link to Funnel.

list_export_connections

Read

Lists the connections you can use for a destination.

prepare_create_export

Write (prepare)

Prepares a new export.

prepare_set_export_enabled

Write (prepare)

Prepares turning an export on or off.

prepare_run_export

Write (prepare)

Prepares running an export now, optionally for a date range you choose.

execute_export_change

Write

Makes the export change you approved.

The prepare_* tools don't change anything. Each approved change runs once.

​

What to expect

The new export appears with your other exports in Funnel and runs every day at the time you set. If the table, files or spreadsheet don't appear, open the export in Funnel and check its run status.

​

You can do the same tasks in Funnel AI, except create Snowflake exports. See Manage exports in Funnel AI.

Did this answer your question?