Death Plan Switch: a free self-hosted dead man's switch for solo developers

What happened

A developer running a one-person web studio in regional New South Wales, Australia, has released Death Plan Switch, a free, MIT-licensed dead man's switch built for independent web developers whose clients would be stranded if they suddenly vanished. The project reached Hacker News's front page on 20 October through a Show HN post, with the source code published on GitHub.

The premise is simple: the software emails its operator once a week. If the operator stops responding, it escalates through a fixed sequence — next of kin first, then the people who need to act — until either someone confirms the operator is fine or the pre-written messages go out.

How the escalation works

According to the project's website, every step is timed from the weekly check-in email, which arrives at the same time each week. Miss it, and a reminder goes to both registered email addresses at the 24-hour mark, followed by a last reminder at 48 hours. At 72 hours the system asks the operator's next of kin whether everything is all right, and repeats the question at 120 hours. At 144 hours, each named contact receives the message the operator wrote for them. Because of the weekly cadence, contacts learn of a problem at most 13 days after the last successful check-in.

Two safeguards shape the design. If next of kin says the operator is well, the process ends and nobody else hears anything. If they say something is wrong, contacts get their messages immediately instead of waiting out the timer. And the operator can check in at any stage — even after the switch has fired — at which point anyone already contacted is told it was a false alarm.

One detail is notably pragmatic: opening the link in the email does not count as a check-in, because automated email scanners routinely fetch links. The operator has to press a button on the page itself.

Why it was built

In the accompanying post, the developer explains that their clients' sites, domains and hosting all run through them personally. Their existing continuity plan is a single encrypted document describing everything the business depends on, with the password held by a backup developer and a second copy held by a lawyer. What was missing, they write, was the piece in between: something that notices when the operator goes quiet and then notifies people in a deliberate order. Finding no simple option they could host themselves, they built one.

The developer has been running it on themselves for the past week, with next of kin and backup developer both confirmed, and is asking anyone who trials it for feedback on rough edges and desired changes.

Deliberate limits

The project is unusually candid about what it is not. It sends email; it cannot know why someone stopped checking in, and it stops working if its own server, scheduler or mail service goes down. It stores no documents and no passwords — messages are meant to point to where things live and whom to call. There is no hosted version and no account system, and communication is email-only, with no SMS, phone calls or app. The MIT licence comes with no warranty.

What it takes to run

Requirements are modest: a web server with PHP 8.1 or newer, an SMTP account, HTTPS, and a cron job or any external scheduler that requests a web address every 15 minutes. There is no database. Apache works as shipped, while nginx needs four lines of configuration. The developer advises hosting the switch separately from the sites it protects, so a single outage cannot take down both, and says installation is done in the browser in a few minutes, with the installer generating the crontab line — the only genuinely technical step.

Why it matters

Solo operators are businesses with a bus factor of one. A single person often holds the keys to clients' domains, hosting and deployments, and most continuity advice stops at writing things down and handing someone a password. That solves custody but not detection: the plan only helps if somebody notices the operator is gone and knows to look. A dead man's switch fills that gap, and a self-hosted, MIT-licensed implementation lets an independent developer adopt one without a subscription or without entrusting their death plan to a third party that may itself disappear. The trade-off is acknowledged rather than hidden — the switch is only as reliable as the infrastructure it runs on. As working code that can be run as-is or adapted, it lowers the barrier to a piece of planning most one-person businesses know they should do and rarely get around to.