Zero-knowledge encryption

Zero-knowledge encryption is a model in which the service provider never has access to the user's unencrypted data. Data is encrypted on the user's side before it reaches the provider's servers, and the encryption key never leaves the user's control. The server stores only encrypted data and, at no point, holds the key or the plaintext — so only the user can decrypt their own data.

The defining property is where encryption and decryption happen:

• Client-side encryption — data is encrypted on the user's device before upload.
• The key stays with the user — derived from a secret the provider never receives.
• The server stores ciphertext only — it can hold and sync data it cannot read.
• Only the user decrypts — plaintext exists only on the user's side.

With standard server-side encryption, the provider encrypts data but also holds the keys, which means it can technically read the data — and so can an attacker who compromises the provider. With zero-knowledge encryption, the provider never holds the key, so it cannot read the data and neither can anyone who breaches its servers. The trade-off is that if the user loses their key or master secret, the data generally cannot be recovered.

A password manager holds an organization's most sensitive secrets, so the question of who could read them matters enormously. Zero-knowledge encryption means that even the password manager's provider cannot see stored credentials, and a breach of the provider yields only unreadable ciphertext. It is one of the strongest trust guarantees a password manager can offer.