Loading...
Preparing your workspace
Loading tool...
Please wait.
Free: up to 0.5MB / 1,000 rows per side in browser. · optional AI Assistant on exceptions · Export CSV when matched
Checking, savings, credit card, or payout settlement export (≤0.5MB)
General ledger, AR/AP register, or spreadsheet book entries (≤0.5MB)
Free: 1 bank + 1 ledger pair in browser. Paid unlocks batch ZIP for up to 5 jobs.
Upload or paste bank statement and ledger CSV files to match transactions, review exceptions in a mailbox view, and download a reconciled CSV. Adjust date and amount tolerances as you go, run batch matching when you have multiple statement pairs, and use the AI Assistant option when you want help classifying fees, timing gaps, or duplicates.
White-Label Platform
Paste a small snippet on your domain, use your branding, and sell to your clients with one subscription and no revenue share.
TOOLGRID API GATEWAY
1-month free trial · $0
Start with a 1-month free trial at $0 — try 400,000 API credits (AI paths locked until paid). Then $19.99/mo.
Try 400,000 API credits for 1 month · $0
Start free trialToolGrid MCP
Connect ChatGPT, Claude, Cursor, and other MCP clients to this catalog — one connection, same Developer API credits as the gateway. Try 400,000 API credits for 1 month · $0 on the Gateway plan.
Connect MCPCommon questions about this tool
Provide a bank statement CSV and an accounting ledger CSV. You can upload files or paste CSV text, then map Date and Amount columns if auto-detect needs a tweak.
You get a mailbox of matched, possible, and unmatched groups plus a downloadable reconciled CSV, with size and row limits shown in the tool.
Use it on unmatched or possible-match rows when you want an optional classification such as bank fee, timing difference, or duplicate—only after you explicitly choose the AI option.
Learn what this tool does, when to use it, and how it fits into your workflow.
Finance Reconciliation Mailbox helps you compare a bank statement CSV with an accounting ledger CSV and see which lines match. You upload or paste both files, run matching, then review results in a mailbox-style list. Matched items, possible matches, and unmatched exceptions stay visible so you can confirm, split, link, or mark a resolution before you export a reconciled CSV.
The problem it solves is simple but costly. Bank feeds and book entries rarely line up one-to-one. Clearing delays, wire fees, batch deposits, duplicate postings, and mistyped amounts create open items that take hours to chase in spreadsheets. This tool is for bookkeepers, accountants, controllers, finance ops staff, freelancers, and founders who need a clear bank-to-ledger check without rebuilding a full ERP reconciliation module.
Free use covers one bank file and one ledger file in your browser within the size and row limits shown in the tool. Larger single pairs and multi-job batch ZIP matching are available on a paid plan. An optional AI Assistant can suggest how to classify tricky exceptions after you ask for help on a row. Core matching and export still work if you never use that option.
Bank reconciliation means proving that what left or entered your bank account also appears in your books. A bank CSV is usually a statement or payout export with posted dates, descriptions, references, and amounts. A ledger CSV is often a general ledger extract, AR/AP register, or spreadsheet of book entries with dates, payees, invoice refs, and debit or credit columns.
People struggle when they do this by hand. Sorting two sheets by date is not enough. One bank deposit may cover several invoices. One ledger bill may clear a day or two later. A bank fee may have no matching book line yet. A transposed digit can look like a different amount. Spreadsheet VLOOKUPs break when descriptions differ or when amounts are stored as text.
This tool treats each side as a list of normalized transactions. It looks for date, amount, description, reference, and account fields. It then runs several matching passes and groups rows into matched, possible, or unmatched buckets. You work the open items like a mailbox: pick a group, inspect both sides, decide what happened, then export.
Common places this work shows up include month-end close, Stripe or processor payout reviews, payroll clearing checks, credit-card statement reviews, and audit prep. Anywhere you must explain why the bank total and the book total are not the same, a structured match view beats scrolling two worksheets side by side.
Month-end operating account close: Export the bank CSV and the GL cash account, match the period, confirm possible clearing lags, and export the exception list for the close checklist.
Processor payout settlement: Match a Stripe or similar payout file to AR invoices. Use many-to-one grouping when one deposit covers several invoices posted separately.
Payroll clearing review: Compare a bank payroll ACH to the payroll expense batch. Mark timing differences when the bank posts a day after the ledger entry.
Fee and wire variance triage: When amounts almost match, review possible fee deductions or transposed digits instead of creating a new phantom invoice.
Duplicate entry cleanup: Spot two ledger rows that both claim the same bank charge, then mark the duplicate and keep one open for investigate.
Multi-entity batch close: Paid users can run several bank and ledger pairs in one batch ZIP when closing more than one account or entity in the same session.
Training and demos: Use Sample Input to show juniors how matched, possible, and unmatched groups look before they touch live client data.
Each CSV row is normalized into a transaction with an ISO date, an amount in cents, and text fields for description, reference, and account. Amounts from a single column keep their sign. Debit and credit columns are combined into one signed amount. Rows without both a usable date and amount are skipped and counted separately.
Matching runs in ordered passes. Pass 1 locks exact amount pairs within one calendar day and scores them with text and reference overlap. Pass 2 allows your date tolerance and amount tolerance while requiring the same amount sign. Pass 3, when many-to-one is enabled, looks for small sets of ledger or bank lines that sum to a remaining opposite-side amount within the date window. Pass 4 marks near misses such as transposed digits or small absolute differences that may be fees, as possible matches for human confirmation.
Summary metrics count matched groups, possible groups, unmatched bank and ledger rows, matched money volume, and net variance between bank and ledger totals. Match rate is based on how many rows landed in matched groups versus the overall population of transactions. Resolution labels you apply are stored on the group and included in export; they do not rewrite the original CSV files.
Batch jobs reuse the same matching rules per pair. When AI Assistant is enabled on batch, only a limited number of unmatched or possible groups per job receive suggestions. Failures on one job do not cancel successful jobs. The ZIP stores each successful job’s reconciliation CSV and a lightweight summary, plus a batch summary file and a failures list when needed.
| Path | Capacity | Main result |
|---|---|---|
| Free single pair | 1 bank + 1 ledger; about 0.5 MB per file; up to 1,000 rows per side; 1,500 combined | Browser match, mailbox review, CSV export |
| Paid single pair | Up to 10 MB per file; up to 25,000 rows per side; 40,000 combined | Server match for larger statements, same mailbox and export flow |
| Paid batch ZIP | 1–5 jobs; up to 5 bank and 5 ledger files; 10 MB each; 40 MB combined | ZIP with per-job reconciliation CSV, summaries, optional AI report, failure notes |
| Mailbox status | What it means |
|---|---|
| Matched | The tool linked bank and ledger sides with strong agreement, or you confirmed/linked them. |
| Possible | Amounts or dates are close enough to review, but you should confirm before treating them as final. |
| Unmatched | No pair was found yet. Investigate timing, fees, missing postings, or mapping issues. |
Export clean CSVs with clear date and amount headers whenever you can. If auto-mapping fails, fix Map Columns before matching again. Start with a narrow date range if files are large. Keep many-to-one on when payout deposits commonly cover several invoices; turn it off when you need strict one-to-one only.
Always review possible matches and fee-like variances yourself. Matching is a helper, not a payment approval. Confirm AI Assistant suggestions before accepting them. Free limits protect everyday browser runs; use paid higher capacity or batch ZIP only when your statements exceed those caps or you have multiple pairs.
Do not upload confidential financial data you are not allowed to process. Invalid files, empty sides, or missing date/amount rows will stop or skip matching for that side. Unpaired files in a multi-file batch are reported as failures rather than silently ignored. Related prep tools in the same document workflow include converting statement tables with PDF to Excel, making scans searchable with OCR PDF, extracting invoice fields with Invoice Email Extractor, or intake of vendor bills with Vendor Invoice Workflow Mailbox before you reconcile cash.
Summary: Upload or paste bank statement and ledger CSV files to match transactions, review exceptions in a mailbox view, and download a reconciled CSV. Adjust date and amount tolerances as you go, run batch matching when you have multiple statement pairs, and use the AI Assistant option when you want help classifying fees, timing gaps, or duplicates.
We’ll add articles and guides here soon. Check back for tips and best practices.