To convert a scanned bank statement PDF to Excel, you need OCR — optical character recognition — plus a layout engine that can rebuild the transaction table from the recognized text. A scanned PDF is a picture of paper, not a document with selectable text, so copy-paste and most free converters fail before they start. On macOS, PDF Bank Statement Converter runs OCR on the image, detects the transaction table, and exports a clean XLSX with dates, descriptions, amounts, and balances in the correct columns.
Native PDF vs scanned PDF — and why it matters
Two files can look identical on screen and be completely different underneath. A native PDF was generated directly from your bank's system. It contains a text layer — each character has a position, a font, and a value. You can select it, search it, and copy it.
A scanned PDF is an image wrapped in a PDF container. Somebody pointed a scanner or phone camera at a printed statement, or the bank archived their old statements as images years ago. There's no text inside — only pixels arranged to look like letters. To a computer, it's indistinguishable from a photograph of a wall.
This is the single biggest reason people get stuck converting older bank statements. The file says .pdf, so users expect it to behave like one. Then they open it, drag-select a row, and the cursor draws a rectangle instead of highlighting text. That rectangle is the tell.
The Preview "select text" test
You can diagnose any statement in ten seconds without leaving macOS.
- Open the PDF in Preview.
- Hover your cursor over a single transaction — a date, a description, an amount.
- Click and drag.
If the text highlights cleanly, line by line, you have a native PDF. Copy-paste might still mangle the columns in Excel (that's a different problem, covered in the copy-paste fix guide), but the text is in there.
If instead your cursor draws a blue rectangle — the same behaviour as selecting a region of a screenshot — the PDF is scanned. There is nothing to highlight. OCR is the only way forward.
Why scanned statements break every copy-paste and most converters
"Export to Excel" features in browser tools and generic PDF apps usually scan the file for a text layer, grab whatever they find, and dump it into cells. When the text layer doesn't exist, they produce an empty spreadsheet or, worse, a spreadsheet with a single image pasted into cell A1.
Even tools that claim OCR often stop after the recognition step. They hand you a wall of text — every word, no structure. For a bank statement, that's useless. A statement's value is in the columns: the date next to the description next to the amount next to the running balance. Lose the column structure and you've traded one unreadable format for another.
Scanned bank statements break generic OCR in specific ways:
- Column edges drift slightly between pages, so a one-size-fits-all template shifts amounts into the wrong column halfway down the document.
- Thin table rules and faint gridlines get lost, leaving the recognizer with no boundary to follow.
- Negative amounts, sometimes printed in red or with parentheses, get read as positive numbers or dropped entirely.
- Multi-line descriptions — a merchant name followed by a city and reference code — get chopped into separate rows.
- Headers repeat on every page, so a naive recognizer produces one "Date / Description / Amount / Balance" row per page.
How PDF Bank Statement Converter handles scanned PDFs
The Mac app runs on-device OCR to extract the text from the image, then runs a transaction-table detector specifically trained for bank statements. That second step is the one most generic OCR tools skip. It's what turns a pile of recognized words into a structured XLSX with the right columns and the right number of rows.
Concretely:
- Headers, footers, marketing, and page numbers are filtered out.
- Repeated column titles across pages are collapsed into a single header row.
- Multi-line descriptions are stitched back into one cell.
- Negative amounts are preserved with their sign, whether the bank prints them with a minus, parentheses, or a red tag.
- Running balances are kept alongside the debit / credit columns so you can reconcile.
Files stay on your device while you select and review them. There are no accounts, no trackers, and your statements are never used to train models.
Scan quality checklist
If you're scanning a paper statement yourself — from an old filing cabinet, a landlord's archive, or your own records — the quality of the scan determines the accuracy of the OCR. Bad input, bad output. The good news is that you don't need a professional scanner.
Aim for:
- 200 to 300 DPI. Below 200 DPI, small digits (the 3 vs the 8 problem) start to get misread. Above 400 DPI you're wasting file size for no accuracy gain.
- Straight edges. Skew under 2 degrees is forgiving. A page scanned at a 15-degree angle confuses even good OCR.
- Strong contrast. Black text on white paper. If the original is yellowed or faded, boost contrast in your scan settings before saving.
- No shadow across the page. The classic iPhone-camera problem — your own hand casts a gradient down one side. Use a desk lamp or shoot in diffuse daylight.
- One page per image. Don't stack two pages in a single scan. Let the app process each page separately.
- No staples. Sounds obvious — but a stapled corner that lifts a row off the bed will produce an unreadable row.
Apple's built-in tools are fine for this. In the Notes app on iPhone, tap the camera icon and choose Scan Documents. In the Files app, the same scanner is available from the three-dot menu. Both flatten perspective, correct color, and export a real PDF rather than a photo. Drop that PDF into iCloud Drive and open it from the Mac app.
If you have a flatbed scanner or an all-in-one printer, set it to grayscale at 300 DPI and save as PDF. That's still the gold standard for long archives.
What the app handles that generic OCR doesn't
The work that matters for a spreadsheet is column recovery. Generic OCR — including the OCR built into Preview's markup tools — will cheerfully give you every word on a page. But a bank statement isn't prose. It's a table. And tables scanned from paper rarely survive line-by-line recognition.
The converter reconstructs columns in three ways that general-purpose tools don't:
- Column boundary inference across pages. Even when the amount column drifts two millimetres left on page 4, the app tracks the running boundary so the amounts don't end up in the description column.
- Transaction-row clustering. A multi-line description ("AMAZON.COM*M3K4L — SEATTLE WA — REF 2241") stays in one row instead of fragmenting into three.
- Amount-sign preservation. Withdrawals and refunds keep their sign whether the bank uses
-125.00,(125.00), or aDR/CRmarker.
The output is an XLSX (or CSV, TSV, Markdown, XML, TXT) where each row is one transaction and each column means what you expect it to mean. If you're handling a stack of scans — months or years of archived statements — the Mac app batch-converts folders in one pass, keeping the column structure consistent across every file.
When to use the Mac app vs the iPhone app
Both apps run OCR with the same accuracy. The Mac app is the better choice for scanned archives because it supports batch processing, side-by-side review (original PDF on one side, extracted table on the other), and every export format including XLSX. Use the iPhone app when you scan a statement on your phone and want a CSV in iCloud Drive before you sit down at a desk.
For a deeper walkthrough of the full statement-to-spreadsheet workflow on Mac, see Convert Bank Statement PDFs to Excel on Mac. For a bank-specific example with scanned checking-account statements, see Chase bank statements to Excel on Mac.
Get the Mac app
Download Bank Statement PDF Converter on the Mac App Store. Drop in a scanned PDF, let the OCR and table detector run, review the rows against the source, and export to XLSX. Free to try, no accounts, runs natively on Apple Silicon and Intel.
FAQ
How do I know if my bank statement PDF is scanned or native?
Open the PDF in Preview on Mac and try to select a single transaction with your cursor. If you get a clean text selection, it's a native PDF with a text layer. If your cursor drags a rectangular marquee instead — the way it would over an image — the PDF is a scan and needs OCR before it can become a spreadsheet.
Why doesn't copy-paste work on a scanned bank statement?
A scanned PDF is just a picture of paper. There is no text inside it, only pixels. Copy-paste fails because there is literally nothing to copy. OCR has to read the image and reconstruct the text before you can get it into Excel.
What scan quality do I need for a bank statement?
Aim for 200 to 300 DPI, straight edges, good contrast, and no shadow across the page. Apple Notes and the iPhone camera's document-scan mode work well if you flatten the paper and hit the frame squarely. Flatbed scans are the most reliable.
Can PDF Bank Statement Converter recover columns from a scanned statement?
Yes. The converter runs OCR and then a transaction-table detector that rebuilds columns across scanned rows — dates on the left, descriptions in the middle, amounts and balances on the right — so the exported XLSX lines up even when the scanned table has slight skew or noise.
Does OCR work on photos of statements from my iPhone?
Yes, as long as the photo is legible. Use Apple's built-in document-scan mode (in the Notes or Files app) rather than a plain camera shot — it flattens perspective and boosts contrast. Save the scan as a PDF and import it into the Mac app.