Sync Bank Transactions to Notion Automatically

Yes. BankSync can only read your financial data. It cannot move money or initiate transfers, change account settings or passwords, access trading features, or apply for financial products. You enter your bank login in the provider’s own secure window — Plaid, Fiskil or SnapTrade, depending on your institution — never in BankSync, and BankSync never sees or stores your bank password.

You pick the cadence on the feed’s Schedule tab, from four modes: Hourly, Daily, Weekly, or Manual. Which automatic cadences you can choose depends on your plan — Free and Starter run weekly, Standard adds daily, and Professional and above unlock hourly. A manual “Sync Now” works on every plan regardless of the schedule. Scheduled runs happen at the top of the hour in your local time zone, and new rows appear in Notion when the source provider has made the data available.

Transactions, balances, investment trades, holdings, open orders and loan data, plus receipts, invoices and documents — whichever of these the accounts you selected and their provider actually support. One feed carries exactly one data type, so to send both transactions and balances from the same accounts to Notion you create two feeds. Investment accounts cannot be mixed with checking or savings accounts in the same feed.

No. BankSync keys every record on the unique identifier the bank’s provider issues and skips records it has already written for that account, so overlapping date ranges, retried jobs and a manual sync minutes before a scheduled one are all safe. Two things to know: two different feeds writing the same account to the same place deduplicate independently, because each feed keeps its own history; and rows you edit or delete yourself are not re-detected. Deliberately re-running a feed from the beginning can create duplicates, so use that sparingly.

The first run backfills the history the bank’s provider will return, and how far back that reaches is set by the market rather than by your plan: Australia up to 7 years (Fiskil (AU)), United States and Canada up to 2 years (Plaid (US)), and United Kingdom up to 90 days (SaltEdge (UK)). Long backfills are processed in 7-day chunks, so a multi-year first sync takes a while, and how far back an individual bank goes still varies by bank and provider. Balances, holdings and loan data are point-in-time and cannot be backfilled at all — that history starts at your first sync and builds up from there.

Disconnect the bank from the Banks tab. If any feed still uses it, remove it from those feeds first. Disconnecting is a soft delete: the bank stays in your list, greyed out with a “Reactivate Bank” option, for about 30 days until the end-of-month cleanup runs. On permanent removal the underlying provider connection is revoked at the source. Data that past syncs already wrote into your Notion is not touched by disconnecting — it stays exactly where it is.

Enrichments fill fields on transactions as they sync. Your own rules and the values BankSync has learned from your corrections are applied first; the AI Categorizer runs last and only fills fields that are still empty by then, so it handles the long tail of merchants nothing else could name. Merchant Cleanup does the same for the merchant name, turning a raw descriptor like “SQ *BLUE BTL COF 0442 OAK” into “Blue Bottle Coffee”. The AI Categorizer is in beta and available on Professional plans and above.

All plans include a 14-day free trial. A card is collected at signup, but it is not charged during the trial, and you can cancel before it ends.

Only what you share. Notion connects via OAuth: you authorize BankSync in a Notion popup and select exactly which pages and databases to share, and BankSync can only ever see those. If a database is missing from the picker in a feed’s Destination tab later, it usually was not shared — re-run the authorization and add it. Re-authorizing does not disturb your existing feeds.

You pick one of your shared databases as the feed’s target, then map bank fields to its properties. Give it property types that match the data: a Date property for transaction dates, a Title or Text property for descriptions, Number properties for debit and credit amounts, and a Select or Multi-select property for categories.