Bank Reconciliation Matcher
Build an internal tool that imports your bank statement and cash ledger, auto-matches the cleared items, and hands you only the exceptions to resolve - then ties out to the penny and locks a signed-off reconciliation.
Import both files, let the agent auto-match and score the cleared items, review only the unmatched exceptions, sign off a reconciliation that ties book balance + outstanding items to the bank balance, and export a CSV/PDF report with a full audit trail.
Before you start
- A bank statement you can export as CSV (date, description, amount)
- A cash-account / GL ledger you can export as CSV (date, memo, amount, reference)
- Free Vercel, Supabase, and Resend accounts (all have free tiers)
- No coding experience required - you'll paste prompts into an AI agent
The problem this kills
Month-end bank reconciliation is the chore that eats a weekend you didn't have. You export the bank statement, export the cash ledger, line them up in a spreadsheet, and start the eyeball game: this $1,240.18 deposit on the bank side is probably those two invoices that posted on the 3rd and the 4th, this card-processor settlement is probably yesterday's batch minus the fee, and that stale check from two months ago still hasn't cleared. You scroll, highlight, and hope you didn't fat-finger a match. When it doesn't tie out, you start over.
The boring truth is that 90% of the lines match themselves - same amount, same day, obvious. The job is really about the 10% that don't: the grouped deposits, the timing differences, the fees, the prior-period outstanding checks. A spreadsheet makes you do all 100% by hand anyway.
This tool flips that. It does the obvious 90% automatically and shows you only the exceptions that actually need a human - so reconciliation becomes a review, not a re-keying marathon.
What you'll build
A private, login-gated web app where you (and only your team) can:
- Import a bank statement CSV and a cash-ledger CSV for a chosen account and period.
- Auto-match cleared items - exact matches, near-date matches within a tolerance you set, and grouped one-to-many matches (one bank deposit = several ledger entries), each with a confidence score and a plain-English reason.
- Carry forward prior-period outstanding items (uncleared checks and deposits in transit) so they're in play this month too.
- Resolve exceptions - the screen shows matched and unmatched items on both sides; you accept the agent's proposals, override them, or mark an item as still outstanding.
- Tie out to the penny - a live reconciliation summary proves book balance + outstanding items = bank balance before you're allowed to finish.
- Sign off and lock the reconciliation, then export a clean CSV and a PDF report.
Every match decision, override, and sign-off is written to an audit trail with who and when.
What's inside the Implementation Plan
- It interviews you first. Before it builds anything, the plan has the AI agent ask you about your bank, your ledger, your CSV column names, your typical and peak line counts, your date-tolerance and grouping rules, and your messiest edge cases (card-processor fees, voided checks, foreign-currency lines). It reads a short tailored spec back to you, you give a thumbs-up, and only then does it build - so you get a tool shaped to how you actually reconcile, not a generic template.
- A step-by-step build with a ready-to-paste prompt at the end of every step.
- The full data model: bank lines, ledger lines, match proposals, exceptions, reconciliations.
- The matching engine - exact, near-date, and grouped - with confidence scoring you can tune.
- A "No API yet?" fallback so the whole thing works from CSV imports today, with no bank or ERP integration required.
- A verification checklist so you know it actually ties out before you trust it.
The governance it includes (this is the point)
This isn't a throwaway script - it's an internal tool you can defend in an audit:
- Login so only your team can open it.
- Row-level security so each organization only ever sees its own accounts and reconciliations.
- A complete audit trail - every auto-match, every human override, every sign-off, with the user and timestamp.
- A hard human-in-the-loop gate - the AI only proposes matches; nothing is final until a person reviews, the summary ties to the penny, and they sign off. A signed-off reconciliation is locked and can't be silently edited.
- Duplicate guards - a fingerprint on every bank line means the same statement can't be imported and reconciled twice.
Who it's for
Bookkeepers, staff accountants, and controllers who reconcile one or more bank accounts every month and are tired of doing it by hand in a spreadsheet. If you can export two CSVs and you want the machine to do the obvious matches while you make the judgment calls, this is for you. No development background needed.
You've got this - paste the first prompt and let the agent interview you.