Scan receipts to a spreadsheet on iPad (2026 workflow)

It’s the end of the month. There are forty-three receipts on your desk — some thermal-printed and already fading, some crumpled, two from Uber that you screenshotted, and a stack from a conference last week with handwritten tip amounts. Your accountant wants a spreadsheet. Your client wants an expense report. You want to be done in under an hour.

The iPad is, for this specific job, the right tool. The camera is good enough to capture thermal paper without a flatbed scanner. The screen is big enough to actually review extracted rows before you commit them. And on iPadOS 26+, on-device AI can read forty-three receipts and produce one Numbers spreadsheet without sending a single image to a cloud service.

This guide is the workflow that gets you from “pile of paper” to “expense report I can email” without a Mac, without a desktop scanner, and without a subscription to a cloud expense tool that wants $12/user/month for the privilege.

Why receipts are the worst kind of “PDF to spreadsheet” problem

If bank statements are hard, receipts are harder. Every receipt is a one-off. Every restaurant prints differently. Every retailer puts the total in a different place. And the surface itself is hostile — thermal paper that fades, paper crumpled in a wallet, photos taken at angles in fluorescent lighting.

Concretely, the failure modes a receipt extractor has to handle:

Multiple totals on one receipt. Subtotal, tax, tip, grand total. The “amount” you want for an expense report is usually the grand total. The “amount” your accountant wants for VAT/sales tax recovery is usually subtotal + tax broken out separately. A naive extractor grabs whichever number it sees first.
Currency ambiguity. A $ in front of a number on a hotel receipt in Tokyo might still be USD (the hotel quoted in dollars) or might be a sloppy template using the wrong symbol for JPY. You want both the original currency and the amount as a number.
Date format chaos. US receipts say 04/17/26. UK receipts say 17/04/26. European retailers say 17.04.2026. Hotels in Asia sometimes say 2026/04/17. Without a hint, no parser gets this right uniformly.
Handwritten tips. A printed restaurant receipt is one thing. A printed restaurant receipt with a handwritten tip and final total is another — the OCR has to read the print, the AI has to recognize the handwriting overlay, and the math has to reconcile.
Photo geometry. A receipt photographed at an angle with a fluorescent reflection is a worst-case for traditional OCR. Apple’s Vision framework on iPad is good at de-skewing and shadow-correcting, but the downstream model still has to be trained on the result.

A tool that solves only the easy case — a single, perfectly photographed restaurant receipt — is useless for an actual end-of-month batch. Solving the messy cases is the entire job.

What the iPad brings to this specific workflow

You can do this on iPhone, and you can do it on Mac. On iPad, three things line up:

The camera + Center Stage. Even an entry-level iPad has a camera good enough for receipt capture, and on iPad Pro the LiDAR-assisted document scan in iPadOS produces near-flatbed quality with a single tap.
Apple Pencil for review and correction. When the extractor gets a tip amount wrong because the handwriting was genuinely ambiguous, fixing it on a 13” screen with a Pencil is faster than tapping on iPhone and faster than wrestling a trackpad on a Mac.
Stage Manager for true two-pane work. Files on the left, the spreadsheet on the right, drag a row across to your accounting app. This is the one Apple workflow paradigm that actually shines on iPad.

If you’re sitting at a desk doing month-end and your iPad is in front of you, this is the right place to do this work — not because the iPhone version is bad (it isn’t, see the iPhone bank-statement workflow), but because reviewing the output is where the time goes, and the iPad’s screen is where reviewing is fastest.

Method 1: ignitai with the iPad camera (the main path)

The full flow, end to end:

Open ignitai on iPad and tap the camera icon in the top right. The capture mode defaults to multi-page document scan — the same Apple Vision pipeline that powers Notes’ built-in scanner, but feeding straight into the extraction pipeline.
Photograph the stack. Hold each receipt flat against a plain surface (a desk works; a knee on a plane does not). The shutter auto-fires when geometry and lighting cross a threshold, so you can rip through 40 receipts in two minutes — feed in, beep, set aside, next one.
Pick the receipt extraction preset. Or write a custom prompt. The shipped preset extracts: date, merchant, category (one of food, lodging, transport, supplies, fuel, other), subtotal, tax, tip, total, currency, payment method (last 4 digits if visible), and notes (anything handwritten or unusual).
Hit Extract. On iPad Pro M1+ running iPadOS 26+, the on-device model handles 40 receipts in under 90 seconds. On older iPads or older iPadOS, the hosted path takes about the same wall-clock time but routes through the network.
Review the table. ignitai shows the extracted rows side-by-side with thumbnails of the source images. Tap any cell to edit; tap the thumbnail to see the full receipt at full size. This is where the iPad screen earns its keep — you can review 40 rows in maybe four minutes.
Export. XLSX, CSV, or Numbers. Save to Files. Drag into your accounting app or email it to your accountant.

For a recurring expense workflow, save the prompt as a preset with the categories your accountant actually uses. The second month is one tap.

Method 2: Apple Notes + Numbers (manual path)

If you don’t want to install anything and you have a small batch:

Open Notes on iPad.
Tap the camera icon → Scan Documents. Capture each receipt; Notes auto-de-skews.
After capturing the batch, tap the scanned PDF → tap and hold the text on a receipt → use the contextual menu’s Live Text features to pull values one at a time.
Open Numbers, paste each value into a row.

This works for under 10 receipts on a slow Sunday. For 40 it’s a non-starter — you’ll be at it for an hour, and you’ll make at least three transcription errors that you’ll only catch when reconciliation breaks.

Live Text on iPadOS is genuinely good for one-off lookups. It’s not built to drive batch table extraction.

Method 3: Continuity Camera + a Mac

If your iPad is your main machine but you have a Mac somewhere on the network, you can use the iPad as a Continuity Camera document scanner into a Mac app. From there, the Mac workflow takes over.

This is a fine path if you’re already at the Mac. If your goal is to never leave the iPad — month-end during a flight, expenses on the couch — it adds context-switching for no gain.

Method 4: Apple Wallet auto-tracking (limits)

Some merchants now push receipt details directly to Apple Wallet for Apple Pay transactions. This is great for the receipts it covers — the data is structured, the merchant is verified, and you don’t have to scan anything.

The catch: it only covers Apple Pay transactions at merchants that have opted in, which is a small fraction of real-world expenses. Conference per-diems, small restaurants, anything in cash, anything from a vendor that hasn’t integrated — none of it. As a primary workflow it doesn’t exist yet. As a complement to scanning, it eliminates 10–20% of the stack.

The fields that make a receipt spreadsheet actually useful

Receipts get scanned for one of three downstream purposes — and the columns you want differ slightly per case:

Personal expense reimbursement at a job. Date, merchant, total, payment method (proof you paid it). Skip subtotal/tax. Add a purpose column for the project code.
Business books / sales-tax recovery. Date, merchant, subtotal, tax (or VAT) as a separate column, total, currency. If you operate in multiple jurisdictions, add a tax_jurisdiction column.
Personal budget tracking. Date, merchant, category, total. Subtotal and tax are noise here.

Pick the schema before you write the extraction prompt. Re-extracting because you forgot the tax column is wasted time. ignitai’s saved presets exist precisely so the schema decision is made once per use case, not once per batch.

Three details that separate “I have a CSV” from “I can hand this to my accountant”

Currency normalization. If your batch has receipts from multiple countries, add an FX-rate column and convert everything to your reporting currency in the spreadsheet, not in the extraction. The extracted original_amount and original_currency are the source of truth; the converted amount is derived. Don’t conflate them.
Date as ISO. Same point as the bank-statement guide — standardize on YYYY-MM-DD immediately. Spreadsheet date math fails silently otherwise.
Source image preserved. Whatever workflow you use, keep the original photograph file with a name that maps to the row (e.g., receipt_2026-04-12_amalfi.jpg). When your accountant asks “what’s this $84 charge from April 12?” you want to be able to surface the photo in two clicks. ignitai writes the source filename automatically; if you’re going manual, do this by hand.

Skip these three and you’ll find out something is wrong six weeks later when reconciliation fails. Five minutes of discipline on extraction day saves an entire afternoon during quarter-close.

The monthly workflow that actually sticks

The reason most people fail at expense tracking isn’t capability; it’s friction. The pile sits because each receipt feels like a 30-second task that costs 30 seconds of resistance. The trick is collapsing the friction to near zero per receipt:

Scan as you receive. Open ignitai, hit camera, photograph, set the paper aside (or recycle it — the photo is the record). Three seconds per receipt at the moment of receipt.
Extract weekly, not monthly. A weekly batch of 10 is psychologically free. A monthly batch of 40 feels like a chore.
Reconcile against the card statement at month-end. Pull the bank-statement extraction, join on date + amount, flag anything that’s on the card but not in the receipts pile. Those are the ones to chase down.
Archive the source images in a dated folder in iCloud Drive or Dropbox. Don’t delete; storage is free and audit-time you’ll want them.

Done weekly with the on-device pipeline, the entire workflow takes maybe ten minutes per week. The first time you do it monthly and it takes ninety, you’ll switch to weekly and not switch back.

When to step outside this workflow

Hundreds of receipts per week (restaurants, retail with audit requirements) push past iPad-based capture into POS-integrated systems. If you’re at that volume, you already know.
Mileage logs are not a receipt scanning problem — they’re a GPS log problem. Use a dedicated mileage app and don’t try to make receipts cover it.
Subscription invoices that arrive monthly via email aren’t receipts; they’re recurring documents. Extract them once, set up a rule, and don’t re-scan them every month.

Bottom line

For a stack of paper receipts at the end of a month, on iPad, that you want as one clean Numbers or Excel spreadsheet: open ignitai, scan the stack with the camera, hit extract, review on the big screen, export. Forty receipts, under ten minutes, on-device on iPadOS 26+.

The iPad is the right device for this specifically because reviewing the extracted rows is where the time goes, and the screen is where reviewing happens fastest. The iPhone version is right when the receipts arrive in your inbox; the Mac version is right when you have a desktop session blocked out for it; the iPad is right when you’re sitting at a desk doing month-end and want one device to do the whole thing.

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