Plaid integration — live bank feeds for reconciliation

Bank reconciliation is the chore that keeps books honest: does what the bank says match what the ledger says? The Plaid integration removes the manual half of it — getting the bank’s side into the system.

What it does

Connect a bank account through Plaid and Oneop pulls transactions and balances on a daily background sync. Imported transactions are deduplicated against what the workspace already holds, then flow into reconciliation.

The matching is genuinely algorithmic: auto-match scores each bank line against ledger entries — amounts within a cent, dates within five days, confidence boosted when the bank reference appears on the ledger entry — and stores a confidence score per suggestion. You confirm, unmatch or exclude individual lines, open a reconciliation, and finalize it.

What flows, in which direction

• Inbound: bank transactions and balances, daily.
• Nothing is written back to the bank — the feed is read-only, as bank feeds should be.

Setup

Bank feeds require Plaid credentials configured on the deployment; without them, the integration quietly does nothing rather than failing loudly — worth checking with your administrator before relying on it. Connect accounts through Plaid’s own consent flow; bank login credentials never touch Oneop.

What it does not do

• No payments, transfers or account writes of any kind.
• No CSV or scanned-statement import path through this integration — Plaid is the working bank-feed path.
• Feed freshness is the daily sync cadence, not real-time.

Related module and plan notes

Bank reconciliation lives in the Finance module, which is present in every workspace; bank reconciliation itself is a Pro-and-above plan surface. Reconciled activity posts against the same double-entry ledger as everything else — one set of books, not a sidecar.