How to Convert a Scanned PDF to a Fillable Form Using OCR

Turning a scanned form into a fillable one is two separate jobs that get run together and should not be confused. OCR makes the text readable. It does not make the form fillable. Understanding that saves a lot of wasted effort with tools that only do half of it.

The two jobs

OCR analyses the image of the page and produces a text layer sitting invisibly behind it. Afterwards the document is searchable, selectable and readable by a screen reader. The page still looks identical.

Adding form fields places AcroForm widgets — text boxes, checkboxes, radio buttons — on top of the page so answers can be typed and stored.

They are independent. You can add working fields to a scan without any OCR at all: the answers are real data even though the background is a picture. And you can OCR a form perfectly and still have nothing to click.

So decide what you actually need:

• Just need people to fill it in? Skip OCR. Add fields.
• Need the text searchable, or accessible? OCR it.
• Need both? OCR first, then add fields — in that order, because some OCR processes rewrite the page and can disturb fields placed beforehand.

Getting OCR right

Output quality is set mostly by input quality, and no amount of post-processing recovers a bad scan.

Scan at 300 DPI. Below that, accuracy falls off sharply. Much above it, you gain little and the file grows.

Scan straight. Skew is one of the biggest causes of recognition errors. Most OCR tools can deskew; feeding them a straight page is better.

Use greyscale or black and white for text documents rather than colour.

Set the right language, including for multilingual documents, or accented characters and unusual letters come out wrong.

Then proofread. OCR output is recognised, not correct, and it fails in a specific way that matters on forms: confusing 0 with O, 1 with l and I, 5 with S, and mangling anything handwritten. On a document where the numbers matter, check them.

This site's editor does not do OCR — use a dedicated OCR tool for this step, then bring the result back for content edits if you need them.

Then add the fields

With the OCR'd file in hand, place the fields:

LibreOffice Draw is the capable free option — open the PDF, show the Form Controls toolbar, and draw fields onto the page.

Adobe Acrobat Pro can scan the page for lines and boxes and propose fields automatically, which on a clean form saves a lot of clicking. Expect to correct its guesses.

The usual rules apply and matter more on a scan, where the underlying lines are fixed: size fields for the longest realistic answer, align them to the printed lines, give every field a meaningful name, give mutually exclusive radio buttons one shared group name, and set the tab order before you distribute it.

This site's editor is not a form builder and does not create AcroForm fields — it edits content and fills forms that already have them.

When it is not worth doing

Sometimes the honest answer is to rebuild the form rather than restore it.

A scan of a photocopy of a fax will fight you at every stage: poor OCR, crooked lines, fields that never quite align. If the form is one you own and will use repeatedly, recreating it in a word processor and exporting a clean, tagged PDF takes less time overall and produces something far better — searchable, accessible, and editable next year when it needs changing.

Restoring a scan makes sense when the layout is legally fixed, when you do not own the form, or when it is a one-off.

Frequently asked questions

Does OCR make a form fillable? No. It makes the text readable. Fields are a separate step.

Can I add fields without OCR? Yes, and they work perfectly. You just will not have searchable text underneath.

Why is my OCR text full of errors? Usually a low-resolution or skewed scan, or the wrong language setting. Rescan at 300 DPI, straight.

Will OCR read handwriting? Standard OCR is built for printed text. Handwriting recognition is a different and much less reliable problem.

Should I OCR before or after adding fields? Before. Some OCR processes rewrite the page and can disturb fields already placed.