1. Processor payouts
CSV needs a date and amount column. Common names like payout_date and net work automatically.
2. Bank deposits
CSV needs a date and amount column. Credits should be positive amounts.
Processor payouts can settle a few days after the customer payment. Exact amounts match first; a payout split across nearby deposits is also matched.
Don't have both exports handy? — four payouts against five deposits, carrying one of each result this tool can report.
Review queue
Where a payout and a nearby deposit almost agree, the What to check column names the gap and whether it looks like a processor fee. It is a lead to verify against your statement, not a match.
Common questions
How close do a payout and a deposit have to be to count as a match?
The same amount to the cent, landing within the date tolerance you set (default 3 days, adjustable up to 30). A payout that settles a few days after the customer payment is expected; the tolerance is there for that gap, not to paper over a real discrepancy.
What if one payout was deposited as several smaller amounts?
The tool also checks whether a payout matches the sum of several nearby bank deposits, searching up to the first 18 eligible deposits per payout to keep the browser work bounded. Each deposit can only explain one payout, so a single deposit is never counted toward two different exceptions. See why one payout shows up as two bank deposits for a worked example, including what happens when the split does not add up exactly.
How is this different from Synder, A2X, or another paid sync tool?
Those connect live to your payment processor and your accounting platform and keep posting transactions automatically, every day, on a paid plan. This workbench has no connection: you choose two CSV files you already exported, it matches them once in your browser, and nothing is kept between runs unless you download the review CSV. See this workbench vs. Synder for the full comparison.
What does the "What to check" column mean?
Where an unmatched payout has a nearby deposit that could explain it, the workbench pairs them and states the gap in money and as a percentage. If that gap falls in the roughly 0.1%–6% range typical of card-processor fees, it says so. Where the amount is exactly right but the deposit landed later than your tolerance allows, it says that instead and names the gap in days, so you can decide whether to widen the tolerance and re-run. Either way this is a lead to verify against your processor statement, not a confirmed match.
Why is a payout unmatched when the deposit is the right amount?
Because the deposit landed outside the date tolerance you set. Banks post a few days after the processor releases a payout, and weekends, holidays, and month-end batches stretch that further. The workbench looks a week past your tolerance when it pairs exceptions, so a row like this reads "same amount, 5 days apart — outside the date tolerance you set" rather than leaving the What to check column empty. Raise the tolerance to cover the gap and run it again, and the pair matches outright.
Does it recognize my processor's CSV format automatically?
It looks for common header names — date, payout_date, deposit_date, net, and amount — which cover typical Stripe, Square, and PayPal exports. If your export uses different column names, rename the header row before uploading.
Is this an accounting system or does it move money?
No. It is a local, browser-only review aid that reads two CSV files you choose and produces an exception list. It does not connect to any payment provider, does not move money, and is not a system of record — download the review CSV if you need to keep the result.
Are my payout and bank files uploaded anywhere?
No. Everything runs in your browser; the files never leave your device. The optional usage counter records only that a reconciliation ran, not any amounts, dates, or file contents.
How is this different from a paid sync tool like Synder?
Synder is a paid subscription that connects to your payment processor and your QuickBooks or Xero account, then posts every transaction and keeps syncing on an ongoing basis. This workbench connects to nothing: you upload two CSV exports you already have and get a one-time review queue, with no account to make, cancel, or pay for. See this workbench vs. Synder for which job fits which situation.