What is a dead man's switch?
A dead man's switch does something automatically when you stop responding. The name comes from trains: a lever the driver had to hold down, which would apply the brakes if he let go.
The idea moved to software, and now it covers a few different things. This page explains how the mechanism works, and what to check before you trust one with anything that matters.
How LastKnow does it
- You write a letter, name who should get it, and arm it. You can attach files.
- You check in on a schedule you choose, from one day to a year: a tap on the Ping page, the link in a reminder email, or a PIN.
- Reminders come before the deadline. Miss it, and a grace period of one to fourteen days starts, with a last notice before anything is sent.
- If you still have not checked in, the people you named get an email with a link to your letter. Emergency stop revokes those links, opened or not.
The core idea
You set a schedule. You confirm you are still around, on that schedule. If you miss it, something happens that you arranged in advance.
The confirming is usually called a check-in. The something is usually sending a message, releasing a file, or notifying a person. The key property is that no action is required for the switch to fire. You have to act to stop it, not to start it.
That inversion is the whole point. It means the mechanism still works when you are not in a position to do anything.
Where they are actually used
- Personal letters and instructions. Someone writes messages for named people and arms them. If the check-ins stop, the messages go out.
- Access handover. Credentials or documents released to a colleague or family member on a trigger.
- Systems and servers. An automated process that halts, alerts, or hands over if the operator stops confirming.
- Journalism and field work. Someone working in a risky place keeps a record that publishes if they go quiet.
What a good implementation gets right
The mechanism is simple to describe and harder to build well. The parts worth judging:
| Property | Why it matters |
|---|---|
| Grace period | A single missed check-in should not trigger anything. People go on holiday, lose a phone, or simply forget. |
| Escalating reminders | You should be nagged well before the deadline, not surprised by it. |
| A way to stop it | If a message goes out by mistake, you need to be able to revoke it before it is read. This is the part most services skip. |
| Multiple check-in routes | If the only way to check in is one app on one device, a lost phone becomes an incident. |
| Clear wording to recipients | The message should say what actually happened, which is that you missed a check-in. Not more than that. |
The failure modes
Three things go wrong with these systems, and none of them is exotic.
- False triggering. The check-in is too easy to miss, or the grace period is too short. Something sends that should not have. This is the most common complaint about poorly designed ones.
- Silent failure. The opposite problem. A reminder email bounces, a cron job dies, and the switch never fires when it should have. The user believes it is armed and it is not.
- Misread by recipients. The message arrives and the person receiving it reads more into it than is true. Good wording prevents this. Vague wording causes real upset.
Both of the first two are worth pressing on when you evaluate a service. Ask what happens if a reminder email fails to deliver, and ask what happens if you ignore the system for a month. If there is no clear answer, the switch is less reliable than it looks.
What this is not
A dead man's switch is not proof of life, and it is not a legal instrument. It knows whether you checked in. That is all it knows, and all it should claim.
LastKnow is built as a check-in with armed letters. If you are looking for a will, that is a separate thing, and you need a qualified lawyer for it. There is more on that in is a digital will legally valid.
Last updated: 2026-10-19