Convert Scanned PDF to Excel on iPhone (no retyping)

The scanned PDF is on your iPhone because the iPhone is where scans happen now. You photographed a stack of invoices with the Notes scanner. A contractor texted you a picture of a quote. Your accountant emailed back last year’s statements as a scanned bundle, and Mail is where you opened it. The table is right there on the screen — you can read every number with your thumb and forefinger — and yet nothing will give it to you as a spreadsheet. Long-press selects a smear of Live Text with the columns gone. The web converter that handled last month’s digital invoice returns an empty sheet. Numbers has nothing to paste.

That is because a scanned PDF is not a document — it is a photograph of a document — and every tool that depends on a PDF’s text layer has nothing to grab. This guide is how to convert a scanned PDF to Excel on iPhone in 2026 without retyping anything and without uploading the pages anywhere: have AI read the page image the way you do, on-device on iOS 26+, and write a real .xlsx with the columns typed and the numbers ready to sum.

Why the iPhone’s built-in tools get you halfway and no further

Apple actually ships very good text recognition. That is precisely what makes this failure confusing:

Live Text reads the characters but discards the table. Long-press a scanned page in Files or Photos and iOS will happily select the text — as lines. The geometry that made it a table (this quantity belongs to that description, this amount to that row) is gone the moment you copy. Paste into Numbers and you get one column of interleaved fragments: dates, then descriptions, then amounts, in whatever order the recognizer walked the page.
The Notes and Files scanners produce great input, not output. Their document camera deskews, crops, and normalizes lighting — which matters later in this guide — but the result is still a photograph in a PDF wrapper. Scanning is capture, not conversion.
The share sheet has no “to Excel” verb. iOS can print, markup, and OCR a scan. It has no built-in path from a scanned table to a spreadsheet, on any iOS version through 26.
Web “OCR to Excel” converters inherit classic OCR’s failure mode. They recover characters with approximate positions, then guess column boundaries from pixel gaps. On a real-world scan — skew from a copier feed, a shadow across a phone photo, multi-line description cells — the guesses collapse. Forty invoice lines come back as a hundred rows of fragments, or as one column of everything. And you uploaded the document to get that result, which is its own problem when the scan is a signed form or a bank statement.

The failure is not your scan and not your phone. The tools are trying to reconstruct a table from geometry, when the job needs something that reads the page.

What actually works: AI that reads the image, on the phone

ignitai approaches a scanned page the way a person does: it looks at the rendered image, recognizes the table as a table — headers, rows, the relationship between a quantity and its line total — and writes that understanding directly into a spreadsheet. There is no separate OCR pass whose errors you inherit; reading the characters and understanding the structure happen together. A scanned statement and a born-digital statement go through the same pipeline and come out as the same clean rows.

The full iPhone flow:

Send the scan to ignitai from the share sheet. From Mail, Files, Messages, or Photos: tap Share → ignitai. (If it isn’t listed, scroll right in the share sheet and tap “Edit Actions” to enable it.) Multi-page scans are fine — a 14-page scanned statement is one file, one job.
Say what you want, in plain English. For a single-table page — one invoice, one receipt, one price list — skip the prompt and let ignitai take the whole thing. For a document with several sections, one sentence picks the table:

“Extract the line items: date, description, quantity, amount. One row per item. Ignore the handwritten notes and the totals box.”

The prompt matters slightly more on scans than on digital PDFs, because paper accumulates clutter — stamps, margin handwriting, a fax header — and naming what you want tells the model what to ignore. Save the prompt as a preset; the second scan is one tap.
Choose Excel (.xlsx) as the output. ignitai sets the column types during extraction: dates become dates, amounts become numbers, currency columns get currency formatting. The sheet opens in Numbers or Excel for iOS ready to compute, not as text you coax into numbers afterward.
Tap Extract. On an iPhone 15 Pro or newer running iOS 26, the extraction runs entirely on-device using Apple’s Foundation Models framework — the page images are read on the phone and the .xlsx is written on the phone. A multi-page scan finishes in seconds. On older hardware or iOS versions, ignitai falls back to a hosted pipeline with documented zero retention.
Review the preview, then save to Files. ignitai shows the extracted rows before saving. Spot-check the first and last rows, then drop the file in iCloud Drive, Dropbox, or OneDrive — Numbers, Excel for iOS, and Google Sheets all open it natively.

Total user time: under a minute for a typical scan.

Capturing paper yourself: do it with the document scanner

Half the scanned PDFs on iPhones were made on that same iPhone, so capture quality is partly in your hands. The rules that make the downstream conversion clean:

Use the Notes or Files document scanner, not a bare camera photo. Open Files → hit the ⋯ menu → Scan Documents (or the equivalent in Notes). The scanner deskews the page, crops to the edges, and normalizes lighting at capture time — the resulting PDF converts noticeably better than a raw photo of the same page.
Flatten the page and fill the frame. A fold or a shadow through the middle of the table is the kind of defect that actually hurts. Skew, within reason, is not.
One document per capture session if the documents are of different kinds. A single PDF with an invoice, a receipt, and a form invites a vaguer prompt than three PDFs each processed with the right one — though a batch of similar documents is exactly what batch mode is for, below.
Receipts specifically have their own workflow. If what you’re photographing is a pocketful of receipts rather than letter-size pages, the receipt-scanning workflow on iPhone covers that capture-to-spreadsheet path end to end, including thermal-paper quirks.

For scans you received rather than made — copier output, accountant bundles, faxes reborn as email attachments — quality is what it is. Flatbed and copier scans at ordinary office resolution are the easy case. What genuinely hurts: text cut off at a margin and thermal paper faded to gray. If you can’t read a number when zoomed in, neither can the model; rescan or re-request that page rather than fighting the output.

Batches: the folder of scans, handled on the phone

Nobody has one scanned page. The real shape of this job is nine photographed invoices from a job site, or a scanned bundle of last year’s statements. On iPhone:

In Files, select all the PDFs (or the one multi-page bundle). Share → ignitai.
Apply one prompt across the batch, or pick your saved preset.
ignitai processes them sequentially and writes a single .xlsx, one row per record, each tagged with a source_file column — so a folder of scanned utility bills becomes one sheet you can pivot by file.
Pages that fail — an upside-down scan, a blank separator sheet — are listed separately, so you re-run only those.

The phone handles the batch sizes phones actually see: a dozen files, a few dozen pages. If the job is a banker’s box — hundreds of files accumulated over years — do it once at a desk with the Mac scanned-PDF workflow, where a folder of sixty files is one drag. Same app, same presets, same pipeline; the only variable is the screen and the drop target.

A worked example: the contractor’s photographed quote

A contractor texts you a photo of a two-page quote — line items, quantities, unit prices, a total — and you want it in a sheet to compare against another bid:

Save the images to Files, or share them straight from Messages → ignitai.
Prompt: “Extract the line items: description, quantity, unit price, line total. One row per item. Skip the header block and the signature area.”
Output: Excel. Tap Extract.
In Numbers, =SUM() the line-total column and compare against the printed total on page two.

If the sum ties, you have a comparable bid sheet in about a minute, built from a text message. Do the same to the second contractor’s PDF — born-digital quotes go through the identical flow — and the comparison is two columns side by side. If you’d rather the output land directly in Apple’s spreadsheet, the PDF-to-Numbers flow on iPhone is the same pattern with a different output pick.

The verification habit: reproduce one printed number

Extraction from a scan deserves one extra notch of skepticism than extraction from a digital file, and the check is one formula: make the spreadsheet reproduce a number printed on the page. Invoice total, statement closing balance, row count — anything the paper asserts, make Excel assert too.

=SUM() under the amount column, compared against the printed total. Match: done.
Mismatch: sort the column and scan the extremes. A misread digit on a faded line is findable in a minute — classic candidates are a 7 read as 1 on thermal paper, or a parenthesized negative “(312.40)” that needed the prompt to say “amounts in parentheses are negative.”

This takes a minute and converts “OCR output I should probably double-check someday” into a sheet you can hand to someone. A spreadsheet built from paper is only as good as the one total you verified — so verify one.

A note on privacy, which is really a note on scans

Scanned documents are disproportionately the sensitive ones — signed contracts, statements, medical bills, tax forms — because sensitive workflows are the ones still running on paper. A web-based “OCR to Excel” service means uploading those exact pages to someone else’s server. On-device extraction on iOS 26+ means the page images are read on your iPhone, the Excel file is written on your iPhone, and nothing leaves the device. The same statement-grade caution behind the iPhone bank-statement workflow applies doubly to scans, because a scan is so often of exactly that kind of document.

Bottom line

A scanned PDF is a photograph, so every tool that depends on a text layer — Live Text copy, Numbers paste, classic web converters — returns fragments or nothing. The workflow that works on an iPhone in 2026: capture with the Notes or Files document scanner (or take the scan you were sent), Share → ignitai, name the table you want in one plain-English sentence, export .xlsx with the types already set, and verify one printed total with =SUM(). On an iPhone 15 Pro or newer on iOS 26, the pages never leave the phone. Use this when the input is copier output, a texted photo of paper, or your own document-scanner captures; use the scanned-PDF workflow on Mac when the job is a whole archive. The steps are the same on both — which is rather the point.

Get ignitai on the App Store — free download, $19.99/mo unlocks unlimited extractions and batch mode after the 3-day trial.