Skip to main content
Bankflow
Back to blog
Integrations7 min readSeptember 10, 2026By the Bankflow Team

How to Export Bank Statement Data to Google Drive

Send parsed transactions straight into a Drive folder as a Google Sheet, CSV, XLSX, or JSON, and understand why Drive works as a destination rather than a source.

Google Drive logo above a three-step diagram: a statement PDF, the parsed transactions, and a Google Sheet written to a Drive folder

What This Integration Actually Does

Connecting Google Drive to Bankflow gives a parsed statement somewhere to go. You upload a PDF, or a workflow picks one up automatically, and when the parse finishes the transactions are written into a Drive folder you chose, in the format you chose, without anyone opening a download dialog.

The result is a folder in Drive that fills itself. Every statement that goes through Bankflow leaves a clean spreadsheet behind it, named consistently, sitting where the rest of your bookkeeping already lives and shared with whoever that folder is already shared with.

There is one thing this integration deliberately does not do, and it is worth understanding before you set it up, because it shapes how you use it.

Why Google Drive Is a Destination, Not a Source

Bankflow connects to Google Drive under a scope called drive.file. That scope grants access to files the application itself created, and to files you explicitly hand it, and nothing else. It cannot list your Drive. It cannot see a folder full of statements you dropped there yesterday.

The scopes that could do that, drive and drive.readonly, are classified by Google as restricted. Restricted scopes require an annual third-party security assessment, renewed every year for as long as the feature exists, and the cost of that assessment is passed on in one way or another by every product that carries one.

So Bankflow made a trade: keep the non-sensitive scope, and offer Google Drive as a place results go rather than a place documents come from. The connection asks for the narrowest access that does the job, which is also the access most people are comfortable granting.

This is not a limitation you have to work around so much as one you plan around. Statements reach Bankflow another way, and Drive receives what comes out.

What Lands in Your Drive Folder

A Drive output can be written in four formats, and the choice is per output rather than per account:

  • Google Sheet: a real, native Google Sheet, not a CSV with a spreadsheet-sounding name. Drive converts the file on arrival, so it opens in Sheets, sorts and filters immediately, and can be referenced by other Sheets formulas. Google Drive is the only destination that offers this.
  • CSV: the plain interchange format, for anything downstream that reads one.
  • XLSX: an Excel workbook, for people who would rather not work in Sheets.
  • JSON: structured records, for a script or an internal tool that is going to read the folder.

Whichever you pick, the columns are the same ones the manual export produces. The export presets you have already configured apply here too, so a file that appears in Drive on a Tuesday morning and one you downloaded by hand on Monday are byte-for-byte identical. There is no second implementation quietly drifting from the first.

If you are parsing invoices and receipts rather than bank statements, the same destination works, and what lands is the line items and totals rather than transactions and balances.

Setting It Up

The whole path takes a couple of minutes:

  • Connect the account. On the connected accounts screen, choose Google Drive and complete the Google consent screen. Bankflow uses OAuth with PKCE, and the token it receives is encrypted at rest and refreshed quietly in the background.
  • Create a workflow. A workflow is one pipeline: where documents come from, what is read out of them, and where the result goes.
  • Add a Drive output. Choose "Write a file", pick the connected Google account, browse to the destination folder, and choose the format. Browsing matters here: you are picking the folder rather than typing a path and hoping it matches.
  • Add more outputs if you want them. Outputs are a list, not a single choice. The same result can go to a Drive folder and to a signed webhook, each with its own destination.

You can connect more than one Google account. A personal Drive and a practice Drive are two separate connections, told apart by the account they belong to, and each output names the account it writes to.

Where the Statements Come From Instead

Since Drive cannot be the input, a workflow needs one of the other three:

  • A watched OneDrive or Dropbox folder. Both of those connectors read as well as write. Drop a PDF in, and it is picked up on the next sweep. See the guides for OneDrive and Dropbox.
  • A forwarding address. Every account gets a private import address. Forward the statement email your bank sends, or set up a rule in your mail client to forward it automatically, and the attachment is parsed the moment it arrives rather than on a schedule. An allow-list controls who is permitted to send to that address.
  • Uploading by hand. Still the fastest way to handle a one-off, and the Drive output applies to it exactly the same.

A common shape is worth stealing: statements arrive by forwarded email, and the finished spreadsheets land in a shared Drive folder the whole finance team already has open. Nobody uploads anything, and nobody downloads anything.

What Happens When a Delivery Fails

Two rules govern this, and both exist because the alternative is worse.

A failed delivery never fails the import. By the time anything is written to Drive, the statement has already been parsed and is already in your ledger. Losing the extra copy in a folder is an inconvenience; throwing away a successful parse over it would be destroying the thing that worked.

A failed delivery is recorded, not swallowed. The workflow history reports parsing and delivery separately, precisely because a file can parse perfectly and never reach the folder. If a write failed, the error is visible and the delivery can be retried. The failure mode this avoids is the quiet one: believing a folder holds a file it does not.

Security and Scope, in Plain Terms

Worth stating plainly, because cloud storage permissions are where most people's caution sits:

  • Bankflow asks for drive.file, which is Google's non-sensitive tier. It reaches files this app created. It does not read your existing Drive.
  • Tokens are encrypted at rest and refreshed in the background. A connection that cannot be renewed tells you so, rather than failing silently on the next run.
  • Revoking access is done from your Google account or by disconnecting inside Bankflow. Either ends it.
  • No bank credentials are involved anywhere in this. You are working with the statement your bank already gave you.

When Google Drive Is the Right Destination

Pick it when the people who need the numbers already live in Google Workspace. Native Sheets output is the deciding feature: the file arrives ready to sort, filter, pivot, and reference from another sheet, with no import step and no conversion. For a team that builds its reporting in Sheets, that is the difference between a file and a working document.

Pick OneDrive or Dropbox instead when you also want the folder to be where statements arrive, since those two read as well as write.

Or use both. A workflow can watch a Dropbox folder for incoming PDFs and write finished Google Sheets into Drive. The connectors are independent, and nothing stops one workflow from spanning two clouds.

If you are new to the parsing side of this, start with how AI bank statement analysis works, then come back and point the output wherever it needs to go. Fifty free page credits, and no card, is enough to try the whole path end to end.

Turn your next statement into usable data.

Review the extracted transactions before export. No bank connection or credit card required.

Analyze a statement free