To extract text from a PDF with OneNote, insert the PDF onto a page as a file printout, right-click a page image and choose to copy text from all pages, then paste into Excel. OneNote's built-in OCR reads scanned statements too. Expect one column of lines that you split into dates, descriptions and amounts afterwards.
Which OneNote versions extract text from a PDF?
The desktop OneNote for Windows extracts text from pictures and PDF printouts. Microsoft's guide to copying text from pictures and file printouts in OneNote lists OneNote for Microsoft 365, OneNote 2024, OneNote 2021 and OneNote 2016 as the versions it covers.
The feature relies on OCR, short for optical character recognition: software that looks at an image of letters and works out which characters they are. OneNote runs it on any picture you put on a page, including each page of a PDF inserted as a printout. That is why it can read a scanned statement, which ordinary copy and paste from a PDF viewer cannot.
Where your copy of OneNote comes from
OneNote is often already installed. It comes with the desktop Office apps, and Microsoft also offers it as a free download for Windows. If you pay for a shared Microsoft 365 plan, each member gets the desktop apps, OneNote included. The cost of Microsoft 365 Family per person shows what that works out to if you are deciding whether to join someone's plan.
What about a Mac?
The steps in this post follow Microsoft's Windows guide. On a Mac, the built-in tools are a better first try. Converting a PDF to Excel on a Mac for free uses Preview and Numbers, both of which ship with the computer.
Inserting a statement as a OneNote file printout
A file printout is OneNote's way of placing a document on a page as a stack of page images, one image per PDF page. It is the step that makes the PDF readable by OneNote's OCR. You insert it from the ribbon, and a ten-page statement becomes ten images on one note page.
Follow these steps in desktop OneNote for Windows:
- Create a new section or page for the statement, so the images do not mix with other notes.
- Click on the blank page where you want the printout to appear.
- Open the Insert tab and choose File Printout.
- Pick the statement PDF and confirm. OneNote places each page of the PDF on the note page as an image.
- Wait until all pages appear. A long statement takes a little longer.
Before you insert: check the PDF type
Try selecting one word in the PDF with your cursor in any PDF reader. If the word highlights, the PDF already contains text and simple copy and paste, or Excel's own import, may give cleaner columns than OCR. If nothing highlights, the PDF is a scan, and OneNote's OCR is the free tool that can read it.
Keep the notebook private
A notebook that syncs to OneDrive puts a copy of your statement in the cloud. Create the page in a notebook you control, and delete the page once you have the text. A statement printout carries your name, address and full transaction history.
Copying text from every printout page in OneNote
Right-click any page image in the printout, and OneNote offers to copy text from that page only or from all pages of the printout. Choose all pages for a full statement. Then click where you want the text, in Excel or on a OneNote page, and press Ctrl+V to paste it.
Microsoft's guide describes both commands for a multi-page printout, and a single Copy Text from Picture command for a lone image. The single-page option is useful when you only need one month's interest line or one disputed charge.
Paste into OneNote first, or straight into Excel?
Paste into a blank OneNote page first if the statement is long. You can read the text there next to the page images and spot obvious mistakes before they reach a spreadsheet. Then select the clean text, copy it again and paste into column A of an empty Excel sheet.
Pasting straight into Excel is fine for a short statement. Click cell A1 before pressing Ctrl+V, so every line lands in its own row of column A.
When the copy command is missing
Microsoft's guide warns that the command may not appear right away for complex or hard-to-read images, and that some results may take up to 24 to 48 hours to become available. If the menu item is greyed out or missing, leave the notebook open and try again later. A clearer scan at a higher resolution also helps.
Why does OneNote text land in a single Excel column?
OneNote text lands in a single Excel column because OCR returns lines of text, not table cells. Each statement line arrives as one string, such as a date, a description and two amounts separated by spaces. Excel puts each string in one cell. You split it into columns with Flash Fill or formulas, and keep the original column for checking.
Here is a worked example with three made-up lines, as they might paste into column A:
| Row | Column A, as pasted (example) |
|---|---|
| 2 | 03/04 CARD PURCHASE EXAMPLE HARDWARE 42.17 1,208.33 |
| 3 | 03/05 DEPOSIT EXAMPLE CLIENT LLC 650.00 1,858.33 |
| 4 | 03/07 CHECK 1042 120.00 1,738.33 |
Splitting with Flash Fill
Flash Fill is an Excel feature that copies a pattern you type by hand. In B2, type 03/04, the date from A2, and press Ctrl+E. Excel fills the dates for the rest of the rows. Repeat in C2 with the description, D2 with the amount and E2 with the balance. Check the filled values. Flash Fill guesses, and a description with a number in it can throw the guess.
The split result for the example looks like this:
| Date | Description | Amount | Balance |
|---|---|---|---|
| 03/04 | CARD PURCHASE EXAMPLE HARDWARE | 42.17 | 1,208.33 |
| 03/05 | DEPOSIT EXAMPLE CLIENT LLC | 650.00 | 1,858.33 |
| 03/07 | CHECK 1042 | 120.00 | 1,738.33 |
Getting the sign of each amount right
The paste loses the column the amount came from, so 42.17 and 650.00 look alike. Use the balance to recover the sign. If the balance went down from the line before, the amount was a withdrawal. In the example, 1,858.33 minus 1,208.33 is 650.00, a deposit, and 1,858.33 minus 1,738.33 is 120.00, a withdrawal. A formula in F3 such as =E3-E2 shows the change, and its sign tells you which way the money moved.
A formula for the last word
If Flash Fill struggles, this classic formula pulls the last space-separated value from a cell, which is the balance on most lines: =TRIM(RIGHT(SUBSTITUTE(A2," ",REPT(" ",100)),100)). Wrap it in VALUE() to turn the text into a number. If you prefer a browser spreadsheet, Google Sheets' free version has a SPLIT function that does a similar job.
OneNote OCR misreads to check on a scanned statement
OneNote OCR misreads the same few things on statements: look-alike characters, decimal points and commas, wrapped descriptions and the page furniture around the table. Microsoft's guide says OCR quality depends on the image and advises looking over the result. On a bank statement, a single wrong digit throws off every balance below it.
Work through this checklist after each paste:
- Look-alike characters. 0 and O, 1 and l, 5 and S, 8 and B. A balance with a letter in it will not convert to a number, which makes these easy to find with a filter on errors.
- Lost decimal points. 42.17 read as 4217 is a hundredfold error. Compare the count of digits after the point on every amount.
- Commas read as points. 1,208.33 read as 1.208.33 breaks the number conversion.
- Wrapped descriptions. A long merchant name that wraps onto a second printed line arrives as a separate row with no date or amount. Join it to the row above.
- Headers and footers. Column titles, page numbers and the bank's address repeat on every page. Delete those rows.
- Dropped minus signs and CR or DR marks. If the bank prints withdrawals with a trailing minus, the OCR may skip it. The balance check that follows catches this.
The balance check
Start from the opening balance printed on the statement, add each amount with its sign, and compare the result with the closing balance. If they match, every amount and sign adds up to the cent, though a misread description or date can still slip through. If they differ, the size of the gap often points to the faulty row, for example a gap of 37.95 when one line reads 42.17 but should be 4.22.
Dates and the books
If you keep cash-basis books, you record income and spending on the day money moves, which is the posting date on the statement. Cash basis accounting for a small business explains why that date, and not the purchase date, is the one to keep in your sheet.
When OneNote is the slow way to extract statement text
OneNote is the slow way when the statement is long, when the PDF already has real text, or when you need the same bank every month. The OCR itself is quick. The time goes on splitting columns, fixing signs and checking digits, and that work repeats for every page and every statement.
Here is how the choice usually falls, by situation:
| Situation | OneNote | Better first choice |
|---|---|---|
| Two-page scanned statement | Good fit, with a page or two of checks | OneNote |
| Text PDF from online banking | Works, but OCR adds misreads | Excel's Get Data import or a copy and paste |
| Twenty-page scanned statement | Twenty pages of splitting and checks | A converter built for statements |
| Same bank every month | The same cleanup every time | A repeatable import or a converter |
Two signs to stop
If the balance check fails on several pages, or Flash Fill keeps guessing wrong because descriptions contain numbers, stop and change method. Each extra pass costs more time than the first one did.
A cheap alternative for text PDFs
For a statement downloaded from online banking, a converter that reads the text layer skips the OCR guesswork. CheapBankConvert is getting ready to open, with early prices of $2.99 for a statement up to 10 pages and free previews of the first 2 pages. It is built for text-based PDF statements, so a scan is still a job for OneNote.