Embedded vs remote signing: which one your flow actually needs
Remote signing emails a link; embedded signing puts the signing page inside your app. What each one costs you, where each one loses people, and how to pick.
Akbar Ali · 25 September 2026
There are two ways to get someone's signature on a document online. The names make them sound more different than they are.
Remote signing: you send them a link, usually by email. They open it whenever, sign on the signing service's page, done.
Embedded signing: the signing page shows up inside your own app, in an iframe. They sign without leaving. Your code gets told when they're done.
Same signature underneath. Same evidence, same certificate, same legal weight. The only thing that changes is how the person gets to the page. That turns out to matter a lot.
The short version
| Remote (emailed link) | Embedded (in your app) | |
|---|---|---|
| Where they sign | The signing service's page | A pane inside your product |
| How they get there | An email, whenever they open it | They're already there |
| Link lifetime | Days. It has to survive "I'll do it Monday" | Minutes. Ours is 30 |
| Code you write | None. Send from a dashboard | An API call and a component |
| Who it's for | People outside your product | Your own users, mid-flow |
| You find out they signed | Webhook or notification | Right away in the page, plus webhook |
Where remote signing loses people
The email.
Everything about remote signing is fine except the part where you hand the next step to someone's inbox. It lands in promotions. It lands next to forty other things. It gets opened on a phone in a lift and "I'll do it properly later". Later is doing a lot of work there.
For a contract going to a client, a supplier, a new hire, that's fine honestly. They're expecting it. They'll chase it. And they're not users of your product anyway, so where else would they sign.
It stops being fine when the signature is a step in your flow. Onboarding, checkout, account setup. Now it's step four of six, and step four says "check your email". You just built a funnel with a hole in it and labelled the hole.
Where embedded signing costs you
Code, mostly. Not much, but it's not zero.
- An API key, and a server call to create the session. The key can't go in the browser.
- A component that mounts the iframe and cleans up after itself. There are a few React traps that are easy to miss.
- An allowlist of the sites allowed to frame the page. Ours ships closed. No origins, no embedding, anywhere.
- Short-lived URLs. That's a feature, but it means you create the session when the user reaches the step, not before.
And one real limit. It only works for people who are in your app. You can't embed a signing page for your customer's lawyer.
So which one
Ask one question. Is the signer your user, and are they in the middle of doing something in your product?
- Yes → embedded. Anything else means sending them away from the exact moment they wanted to be there.
- No → remote. An email link is the right tool for someone who doesn't have an account with you and never will.
Most real flows end up with both. Your customer signs embedded, in onboarding. Then someone on your side countersigns from a normal emailed link, because they're not sitting in your customer onboarding flow, obviously.
The part that doesn't change
People sometimes worry an embedded signature is a lesser one. Like it counts less because it happened inside someone else's page.
It doesn't. It's the same signing session. The same timestamps, the same IP and device captured, the same tamper-evident certificate in the finished PDF. Anyone can check a signed PDF either way and see who signed and when. The delivery is the only difference.
Pick by where the signer is, not by how serious the document is.
If embedded is the one, here's how ours works, and the React guide walks through the whole integration. If it's remote, you don't need any of this. Upload the PDF and send it.
What's in your flow right now? Anybody still doing "check your email" in the middle of onboarding and not minding it? Genuinely curious whether anyone's actually measured the drop-off on that step.