How to sign a PDF without printing or scanning

The contract arrives for approval, the signature is needed today, and there is no printer within reach. The classic route, print it, sign it, scan it, send it back, eats half an hour and damages the document: what comes out is a picture instead of text, the file triples in weight, and search across it stops working.

An electronic signature closes the same task in a minute and, more importantly, does not turn the document into a photograph.

What actually goes into the document

Two things get confused here, and they are worth separating.

The first is the visible mark on the page: the signer's name, the reason, the date. That exists for the person who opens the file and wants to see with their own eyes who signed and when.

The second is the cryptographic signature itself, embedded in the PDF structure. A human never sees it, but software checks it. That is what answers the question of whether the document changed after signing.

The service applies both. The signature is real, PKCS#7 in the PAdES profile, not a rectangle of text drawn over the page. The difference is fundamental: a picture can be faked in any editor in a minute, while a cryptographic signature breaks if a single byte of the file changes.

Why Acrobat will not show a green tick

It is fairer to say this up front, because the result will otherwise puzzle you.

The signature is applied with the service's own certificate, and that certificate is self-signed. It does not build a chain of trust up to a root authority that Acrobat or your corporate system recognises. So on opening the document you will see not a green "signature is valid" tick but a message along the lines of "the signer's identity has not been verified".

What works and what does not:

This is exactly the signature that suits internal approvals, acts between familiar counterparties, sending a layout for sign-off. It does not replace a qualified electronic signature where the law demands precisely that: filing statutory reports, bidding, registration actions. Those need a certificate from an accredited authority, and no online tool can issue one.

The signature fields

Three text fields get filled in, and all three land in the signature properties, not only on the visible mark.

Signer name. A technical placeholder is filled in by default, so it has to be replaced. If the document goes outside, use the full name in the form it appears in the contract itself, not an abbreviation.

Reason. The default is a neutral "Approved". The field carries meaning: it says what the signature is meant to express. "Approved with comments" and "Approved for release" are different things, and the recipient benefits from seeing which.

Location. Optional, a city or an address. It gets left empty more often than filled, but for documents where jurisdiction matters it saves a separate question.

Where the mark lands

There is a limit worth knowing up front: the mark always lands on the last page of the document. You cannot pick page three of fourteen. What you set is only the position within that last page, and you set it on the preview rather than by numbers.

Within the page, put it where a signature is expected: in the details block, next to the space left for signing. A mark in the middle of the text looks careless and sometimes covers content.

The mark itself is not a picture but real selectable text laid over the page without destroying it. The date is filled in by the service, you do not type it. The same name, reason and location are also written into the document properties, so software can read them and not only human eyes. The original author and subject of the document are left alone rather than overwritten.

If the document has several places for signatures, remember the file still carries one cryptographic signature covering the whole document. Several visible marks do not mean several signatures, and placing them in different spots is not something the tool offers anyway.

Long text gets truncated

A practical point that trips people up with long organisation names.

The visible mark is a block of limited size, and text that does not fit is cut. The service warns you that the caption was shortened, so it will not slip by unnoticed, but it is better not to get there.

The reason field suffers most, because it invites a full formulation. "Approved as to the scope of work, except for clause 4.2" will not fit in the mark. If the caveat matters, it belongs in the document text or the covering letter, and the signature reason should carry a short formula.

The same happens with names of long legal entities. Here it helps that the name lands not only on the picture but in the signature properties: in the signatures panel it is visible in full even when shortened on the page.

If a truncation warning arrives and you would rather not change the text, enlarge the mark area when placing it. An empty details block usually has room.

If the document has fields to fill in

Order matters here, and it is easy to get backwards.

Fill the fields first, sign second. Done the other way round, filling the fields counts as changing an already-signed document, and the signature stops being valid.

Fields are filled with pdf-filler, and only once every value is in place and checked does the document go for signing. The same applies to any other edit: the signature is always the last operation.

Step by step

1. Open sign-pdf and upload the document. 2. Replace the signer name with the real one. 3. Write a meaningful reason instead of the default. 4. Place the mark in the details block. 5. Download the result and open it to check both the look of the mark and the signatures panel.

Step five is not a formality. The signatures panel is exactly where you can see whether the cryptographic signature landed or the file kept only a visible mark.

If the signature did not land

That is possible, and the service says so plainly rather than pretending everything is fine.

When cryptographic signing is unavailable for whatever reason, the document still comes back with a visible mark, but with an honest warning: the mark is there, the document was not signed cryptographically, and its integrity cannot be verified.

Take that seriously. Such a file looks signed but is in substance an ordinary PDF with a picture on it. For internal correspondence that may pass; for sending to a counterparty it does not.

What not to do

Do not print and scan. A printed and scanned document loses its text layer: search does not work across it, you cannot copy details out of it, and it weighs several times more. Meanwhile a photograph of a signature carries no more evidentiary weight than an electronic mark, and often less.

Do not paste a signature image in through a photo editor. It is the most common approach and gives the worst possible result: the document looks signed while no verifiable link exists between the signature and the content.

Do not sign a document that will still change. Any edit after signing invalidates it, and you will be signing again. Final version first, signature second, and if the file also has to be protected from changes, protect-pdf goes on top of the signature.