Signing a PDF: what a drawn signature does and does not prove

Two quite different things are called "signing a PDF", and conflating them leads people to over-trust one and over-engineer the other.

Stamping a visible signature

This places an image of your signature onto the page at a chosen position. The result is a PDF that looks like a signed document.

What it establishes: that somebody placed a signature-shaped mark on the page. That is all. There is no cryptographic link to your identity, no record of when it happened, and no way to detect whether the document was altered afterwards.

For most of what people sign in a year — an internal form, a school permission slip, a delivery acknowledgement, a supplier acceptance — this is exactly proportionate. Both parties know each other, nothing is contested, and the signature is a formality that closes a loop.

Sign PDF does this in the browser: draw with a mouse, trackpad or finger, choose the page and position, and download. The signature is embedded as a transparent image using pdf-lib, and neither your document nor your signature leaves the device.

Certificate-based digital signatures

This is a different mechanism. A cryptographic key, bound to a verified identity by a certificate authority, produces a signature over a hash of the document's contents.

What it establishes:

• Integrity — any change to the document after signing invalidates the signature, and viewers say so.
• Authenticity — the signature is tied to a certificate issued to a verified identity.
• Non-repudiation — it is hard for a signer to plausibly deny having signed.
• Timing — with a trusted timestamp, when the signature was applied.

This requires a certificate, and issuance requires identity verification, which is why it involves a provider rather than a browser tool.

Choosing between them

Use a visible signature when the document is routine, the parties know each other, and nobody expects it to be litigated.

Use certificate-based signing when the document carries real financial or legal weight, when multiple parties must sign in sequence, when an audit trail is required, when a regulator or counterparty specifies it, or when the signature must remain verifiable for years.

If you are unsure which category you are in, that uncertainty itself suggests the second.

Practical points for visible signatures

Draw it once and keep it. A signature drawn with a trackpad looks like a signature drawn with a trackpad. Signing on paper, photographing it, and cropping to a transparent PNG gives a much better result — and you reuse it.

Mind the resolution. A signature captured at low resolution and scaled up on the page looks pixellated and cheap. Capture larger than you need.

Position deliberately. A signature overlapping printed text or spilling outside the signature block reads as careless. Adjust before downloading, not after.

Sign the final version. Any later edit means re-signing, and a signature applied to a superseded draft is worse than none.

Check what else is in the file. A signed document is a document you are sending onward — the metadata, annotations and hidden content described in what is hidden inside a PDF go with it.

The reusability problem

Because a visible signature is just an image, anyone who receives the signed PDF can extract it and place it on another document. There is no technical defence within the format; the protection is contextual, in who receives your documents and what a forged instance would have to survive. It is another reason that anything genuinely valuable warrants the certificate-based route.