A message that destroys itself after being read

The written word has an awkward property: it cannot be unsaid. Whatever you send exists, gets copied into backups, syncs across devices and waits patiently in a history. Usually that is useful. Sometimes it is exactly wrong: the thing was meant for one pair of eyes and one moment, not for an archive.

That is what a self-destructing message is for: a message that can be read exactly once. Then it stops existing.

How it works

1. On Secret message, write what you need to pass on.
2. The message is encrypted in your browser and becomes a link. You choose how long it may wait to be read: a day, a week, a month.
3. Send the link any way you like: email, chat, SMS.
4. The recipient opens it, clicks the button and reads the message. That moment destroys it: the server hands the message over and deletes it in the same motion. A second visit only says the message no longer exists.

An unread message does not wait forever; it deletes itself when the chosen expiry passes.

How this differs from disappearing chat messages

You know disappearing messages from Snapchat or WhatsApp. The difference is who you have to trust that the message really disappeared.

With a chat, you trust the app and its operator: your message sits readable on their servers the whole time, and "disappearing" is a promise made by the interface. A secret message is built differently. It is encrypted in your browser, and the decryption key travels only in the part of the link after the hash sign, which browsers never send to any server. The server holds ciphertext it cannot open and deletes it on first release. You do not have to trust a promise; it is enough that nobody but the link's holders ever had the key.

The second difference: disappearing messages work only between users of the same app. A link works anywhere text can be sent.

What it is good for

Passwords have their own article; this is about everything else:

• Personal things. A message meant for one person that should live in no history: an apology, an admission, something you do not want to negotiate with an archive.
• Surprises. The code to a locked cabinet with a present, the meeting point for a surprise party, the wording of a toast. After reading, no trace remains for the guest of honour to stumble on.
• One-time work information. The office door code for a contractor, a parcel pickup PIN, temporary access for a seasonal hire.
• Dictating sensitive details. An ID number for a contract, a document number for a reception desk. Dictating over the phone is a pain; sending it into a chat forever is a mistake.

A rule of thumb: if it would bother you that someone reads this message five years from now, it should not be sent in a way that survives five years.

Frequently asked questions

Will the sender know I read the message?

Not directly; there is no read receipt. Indirectly yes: if they open the link again, they will see the message no longer exists.

What if someone other than the recipient reads it?

Whoever holds the link can read it, so protect the link as you would protect the message. But the one-time property works in your favour: if an intruder claims the message, the intended reader finds it gone and everyone knows immediately that something is off. An ordinary message can be read silently, forever.

Can I destroy the message before anyone reads it?

A message dies by being read or by expiring. If you need certainty sooner, open the link yourself; that destroys it.

Can you read the messages?

No. Only ciphertext reaches us; the key stays in the part of the link browsers never send to servers. Messages are tied to no account, and we do not know who sent them.

What does it cost?

Nothing. The tool is free, with no account needed.