Scanned PDFs are pictures of pages, so normal converters see nothing to copy. MultiPDFsToExcel reads them with an AI vision model and pulls out the fields you ask for.
Start extracting freeWatch the 20-second demoFree to start · No credit card · Sign in with Google
You choose the columns. Each scanned document becomes one row. Here is an example of the output:
| Document date | Reference # | Party name | Amount | Signed (Yes/No) |
|---|---|---|---|---|
| Jan 1, 2019 | AG-0042 | Harbor Freight Lines | $50,000 | Yes |
Example row for illustration. Your columns are whatever you name.
If you can't highlight the text in a PDF, it is an image. Traditional OCR turns those pages into garbled text, especially when the scan is faded, rotated or old, and you end up retyping it anyway.
Old archives, faxed forms, photographed pages. Mix them with digital PDFs if you like.
Name each value you want and add a hint if it helps, like 'the total at the bottom of page 1'.
Values come back cleaned up, for example '$50,000' instead of the surrounding sentence, with a confidence score for each.
Plain OCR only turns pixels into text. Here an AI model reads the page image, understands it, and returns just the value you asked for.
It is built for them: faded print, skewed pages and common OCR mix-ups like O for 0 are handled, and low-confidence values are flagged for review.
It uses the text where it exists and switches to vision for image-only pages automatically.
Upload your first batch and download the spreadsheet in minutes.