Sign a PDF →

Guides Workflow

Make a Scanned PDF Searchable, Then Sign It

Someone emailed you a contract that started its life on paper, got scanned to a PDF, and now sits in your inbox as a pile of images with no selectable text. You can read it, but you cannot search it, paste from it, screen-read it, or feed it to a summariser. And you need to sign it before close of business.

The fix is two tools in a row: OCR to make the scan searchable, then signing to apply your signature on top of the now-searchable file. Both run in your browser. The scan never uploads anywhere — it stays in your tab’s memory the whole time, even during the OCR step (which most online tools do server-side).

Why this chain, in this order

OCR has to come first because the signing tool, like the scan, treats the file as a fixed visual layout. If you sign first and OCR second, the OCR step re-saves the PDF and may move your signature placement subtly. Doing OCR first lets you sign the final, searchable version once, with the signature placed exactly where you want it.

There is a third optional step — compress — for when the OCR-augmented file is larger than the original (it usually is, because the text layer adds bytes). For most scans the size delta is small enough to skip the compression step.

Step 1: Run OCR on the scan

Open the OCR PDF tool and drop in the scanned PDF. Signegy uses Tesseract.js, a JavaScript port of Google’s Tesseract OCR engine, lazy-loaded into your browser only when you click run. The model file is around 30 MB on first use and gets cached for subsequent runs.

Pick the language (English, French, Spanish, German, etc.). The processing happens locally — your CPU does the work. Expect roughly 2-8 seconds per page for a typical letter-sized document on a modern laptop, longer on a phone or older machine. You’ll see a progress bar; do not switch tabs during processing because OCR yields better results when it has a foreground tab to schedule on.

When it finishes, download the searchable PDF. The visible scan looks identical, but underneath there is now a text layer you can search, select, and copy from.

Step 2: Sign the searchable PDF

Open the Signing tool in a new tab and drop in the OCR’d file. Add your signature on the appropriate signature line — drawn with your trackpad or stylus, typed in a script font, or uploaded from an image of your handwritten signature. Place a date stamp if the document calls for one.

Because the file is now searchable, you can use Ctrl/Cmd-F to jump straight to the signature line if the contract is long. That is a quality-of-life improvement that makes signing faster on the third or fourth iteration of a back-and-forth.

The signing tool produces a free audit certificate alongside the signed PDF, recording a SHA-256 hash of the final document. This is useful if you need to prove later that the signed scan has not been altered, but it is not required for the file to be valid.

Step 3 (optional): Compress

OCR adds a text layer to every page, which adds bytes. If your final file is comfortably under your recipient’s email cap (most cap at 10 MB), skip this step. If it is over, run the file through the Compress tool. The text layer takes very little space, so the size reduction usually comes from re-encoding the underlying image at slightly lower quality. Confirm the visible content is still legible before sending.

What to watch out for

Tesseract.js is good but not perfect. Two things consistently trip it up: handwritten text (it is trained on print) and very low-resolution scans (under about 200 DPI). For both cases the text layer it produces will have errors — words misread, punctuation dropped, line breaks in the wrong place. The visible scan still looks correct (because the original image is unchanged), but searches will miss things and screen readers will mispronounce. If accuracy matters — for example, you are running this on a contract that will be quoted in litigation — spot-check the text layer by selecting a few paragraphs and comparing against the visible scan before you rely on it.

The signing step does not re-OCR. If you sign the file and then later realise the OCR was wrong, you have to re-run OCR on the original scan, sign the corrected output, and re-issue.

Signegy provides general information, not legal advice. Confirm any specific evidentiary or filing requirements with your matter’s protective order, court rules, or supervising counsel.