Convert PDF to Excel
Pull the tables and fields out of a PDF into a spreadsheet. Name the fields once and every similar document comes back in the same shape.
- A row per line, in the columns you name
- A whole folder lands in one sheet
- Scans and phone photos work too
50 free tokens when you sign up. No card needed.
Want to try it first? One PDF free, no account.Extract data from PDF
| Date | Description | Money out (debit) | Money in (credit) | Balance |
|---|---|---|---|---|
| 2026-03-01 | Start balance | — | — | 1284.55 |
| 2026-03-02 | Direct Debit to City Water Utilities Ref 123456789 | 42.18 | — | 1242.37 |
| 2026-03-03 | Card Payment to Highstreet Supermarket on 1 Mar | 63.42 | — | 1178.95 |
| 2026-03-05 | Received From A Sample Client Ref Invoice 123456 | — | 480 | 1658.95 |
| 2026-03-08 | Card Payment to Metro Travel on 6 Mar | 18.6 | — | 1640.35 |
| 2026-03-11 | Direct Debit to National Energy Co Ref 987654321 | 88 | — | 1552.35 |
| 2026-03-14 | Bill Payment to London Borough Council Council Tax 2026/27 Instalment 12 Ref 123456789 | 176 | — | 1376.35 |
| 2026-03-17 | Card Payment to Online Marketplace Ltd on 15 Mar | 29.99 | — | 1346.36 |
| 2026-03-20 | Faster Payment From Example Client Ltd Ref March Retainer | — | 1250 | 2596.36 |
| 2026-03-22 | Standing Order to J Smith Ref Savings | 250 | — | 2346.36 |
| 2026-03-25 | Card Payment to Corner Coffee Co 421 on 24 Mar | 8.85 | — | 2337.51 |
| 2026-03-27 | Direct Debit to Mobile Network Ltd Ref 987654321 | 34 | — | 2303.51 |
| 2026-03-28 | Cash Withdrawal at Baker Street on 28 Mar | 60 | — | 2243.51 |
| 2026-03-31 | Interest Paid | — | 1.12 | 2244.63 |
| 2026-03-31 | End balance | — | — | 2244.63 |
Real PDFs, and the spreadsheets they produced
Every one of these came out of the product. Pick a document like yours and compare the PDF against the output it returned.
Why PDF to Excel conversions come out scrambled
A PDF stores text at positions on a page, not values in cells, and a scan stores only a picture. Most converters guess the grid from where the text happens to sit, so a description that wraps becomes two rows, a column of amounts slides out of line with its dates, and page headers land in the middle of the data. Everything looks present until you try to add it up.
We read the table as a table. Each line of a statement or invoice becomes one row, each value lands in the column you named, and a description that wraps stays in one cell. Where the document prints its own totals, a check can compare the rows against them, so a missed line shows up before you rely on the sheet.
A row per line
Every line of a statement or invoice becomes its own row, with the columns you named, ready to sort and total.
A batch lands in one sheet
Send a folder of documents and the rows merge into a single spreadsheet rather than one file each, as Excel or CSV.
You choose the fields
Name what matters once as a template, and every document after that comes back in the same shape.
Before you ask
What can it do?
Two things. It rebuilds a PDF as an editable Word document with the layout intact, or it pulls the specific fields and tables you name into Excel, JSON or Word.
Does it work on scanned PDFs?
Yes. Scans are what it is built for. A photographed or scanned page goes through OCR and comes back as text you can edit, with the page still laid out the way it was. Read the guide.
How do I convert a scanned PDF to Word?
Upload it and download the Word file. Each page is read and rebuilt with its headings, columns and tables, and you get back a .docx you can edit. One page is free with no account.
Will my tables survive?
Yes, as real Word tables with rows and columns you can edit, not a picture of a table and not a stack of floating text boxes.
Can you translate it as well?
Yes, in the same run. The translation lands in the same structure as the original, so you get a translated document rather than a translated wall of text.
Does it read handwriting?
Yes, handwritten pages are supported.
Is the output real text, or a picture of text?
Real text. Every word can be edited, searched, copied and run through a spellchecker, the same as anything you typed yourself.
What does a page cost?
Document conversions start at 9p a page and data extraction starts at 22p a page. The price of a token falls as the plan gets bigger.
Which languages does it work with?
33 languages written left to right, most of them European, including English, French, German, Spanish, Polish, Romanian, Italian and Dutch. Chinese, Japanese and Arabic are not supported yet.
Can I try it before paying?
Yes. You get 50 free tokens when you sign up, no card needed, which is enough to run a few real documents through and see what comes back.
Do I need a subscription?
No. A subscription works out cheaper per page, but you can buy tokens as you go and use them whenever.
Do unused tokens expire?
Monthly allowance rolls over one month, so a slow month is never wasted. Purchased top-ups do not expire.
Are my files saved?
Your original document is deleted as soon as the job finishes. The result is kept for 10 minutes by default, and you choose how long that window is in your settings.
Are my documents used to train AI models?
No. Nothing you upload is used to train anything, ours or anyone else's.
Why does PDF to Word usually wreck the formatting?
Because most converters read the characters first and treat the layout as an afterthought. Columns, tables and headings are a separate problem from the letters, and a tool that solves only the letters hands you a wall of text to rebuild.
Why pay, when other converters are cheaper?
You might save a few pence on the file. You will lose hours putting the formatting back. The saving is your time, not the processing.
Why not just use a generic AI assistant?
When a generic assistant has finished with your file, you still have work to do. It reads a document well and hands back text. We hand back the document, laid out the way it arrived, so you are closer to finished. See for yourself.
Try it on your own document
50 free tokens when you sign up, enough for a few documents before you decide anything.