Skip to content

Connect banks

Connect pulls transactions from banks through Plaid and writes CSV files into your book (Git repository) or into a Connect-only repository. BeanHub then runs the same import rules as a manual CSV upload.

Marketing copy: 12,000+ institutions in 17 countries. Country support is whatever Plaid Link offers in the Connect UI. If a country is not supported, the connect button returns an error modal.

Connect is not on the free plan. Account Settings shows connected-bank usage as unavailable on free. The pricing page lists 10 connected banks on Pro (30-day trial) and writes that as “per repo”. The running app counts connections across the whole account. If those disagree, trust the Connect Banks page in the app. Enterprise is custom.

  1. Open your Git repository.
  2. ImportConnect Banks.
  3. Pick a country.
  4. Click Connect Bank.
  5. Complete the Plaid window (institution, login, accounts).
Connected Banks page with Connect Bank and a country selector.

BeanHub stores the item and starts syncing. Default is a scheduled pull. Edit changes the schedule and the CSV output path under import-data/connect/....

Edit bank connection: schedule and output path.

After CSVs land, import rules that match those paths generate Beancount transactions. Unmatched rows go to Open TXNs.

If you keep Beancount files on your laptop:

  1. Create a Connect Repository (not a Git book).
  2. Connect banks the same way.
  3. Install beanhub-cli (pip install "beanhub-cli>=3.0.0").
  4. bh login (browser grant of an access token).
  5. bh connect sync then bh connect dump, or bh connect dump --sync.

Academy walkthrough with screenshots: Pull CSV with BeanHub’s Direct Connect. Blog: Direct Connect repository.

  • A paid subscription (or trial).
  • Quota: connected banks are counted across the whole account, not per repo. Hitting the cap shows an upgrade message.
  • If a subscription lapses, connections pause, then BeanHub can remove them after about 7 days.

Bank credentials go to Plaid, not into your Git repo. BeanHub stores Plaid access data encrypted. CSV files that land in a Git book are in Git. Treat import-data/ as financial data. Do not make that book public.