Encrypt text

How it works

Encrypting something on a website is normally self-defeating: you hand the plaintext and often the passphrase to a server, and simply have to trust what happens next. This runs in the page, using the browser's own cryptography, so neither ever leaves your device.

What it uses

AES-256 in GCM mode, through the browser's built-in Web Crypto implementation — the same primitives that protect HTTPS, not a hand-rolled cipher. The key is derived from your passphrase with PBKDF2-SHA256 over 250,000 iterations, which makes guessing the passphrase slow for an attacker. A fresh random salt and nonce are generated for every message, so encrypting the same text twice produces different output.

GCM verifies as well as hides

GCM is an authenticated mode: decryption fails if even one character of the encrypted block has been altered. That means it tells you the message is intact, not just that it is secret.

Your passphrase is everything

There is no recovery, no reset and no back door. If you lose the passphrase the message is gone permanently, and that is a property of real encryption rather than a limitation here. The strength of the whole thing rests on the passphrase — a long one of several unrelated words is far better than a short complicated one.

What this is and is not for

It suits sending a note through an untrusted channel, or keeping something in a file you would rather was unreadable. It is not a messaging system: you still have to get the passphrase to the recipient some other way, and doing that badly undoes everything.