Zero-Knowledge Encryption, Explained Without the Marketing Speak

Quick answer: zero-knowledge encryption means the company storing your data holds only ciphertext. Your data is encrypted on your device with a key only you control, so the provider could not read it even if it wanted to. It is a strong, checkable property. Plenty of apps claim it, so check before you trust it.

What does “zero knowledge” actually mean?

In product terms, zero-knowledge means the server has no usable knowledge of your plaintext. It stores bytes that look like noise and never receives the key that would turn them back into your notes. The term is borrowed loosely from zero-knowledge proofs in cryptography, which are a different, more formal idea.

The mechanism is simple. Your password goes through a key derivation function such as PBKDF2 or Argon2id, or your device generates a random key that only your devices hold. That key encrypts your data with a cipher such as AES-256 or XChaCha20-Poly1305 on your device. Only the ciphertext and some non-secret parameters are uploaded. On your other devices the same key is recovered and decryption happens locally.

The server is a mailbox. All the cryptography happens on your devices, once to encrypt and once to decrypt.

How do you verify a zero-knowledge claim?

Check how password reset works, whether a web login shows your data without your password, whether the vendor names its cipher and key derivation, and whether the code is open or audited. A real zero-knowledge service passes all four. A marketing claim usually fails the first two within a minute.

The sharpest test is password reset. If the provider can email you a link, reset your password and still hand your data back, it holds a key somewhere. A real zero-knowledge service replaces that with a recovery key or recovery phrase generated at setup that only you hold.

Second test: a web reader that shows your notes after a magic-link login. If you can read your data without entering the password that protects the key, the server decrypted it, which means the server had the key.

What are the tradeoffs nobody mentions?

Zero-knowledge costs convenience. Forget your password and lose your recovery key, and the data is gone for good. Web access, sharing and search all get harder because the server cannot help. Vendors that hide these tradeoffs are part of why the term has become a marketing word.

• No rescue from support. Nobody can recover the data for you. This is the trade.
• No web preview, or a heavier one. Decryption has to run in your browser, so many services skip the web entirely.
• Harder collaboration. Sharing a note means re-encrypting it to a key the other person can use, usually with public-key cryptography.
• More setup. You have to store a recovery key somewhere safe. Many people do not, and later lose access.
• Search runs on your device. The server cannot index what it cannot read, so large libraries search locally after decryption.

A vendor that is candid about these costs is more likely to be truthful about the rest. If everything is presented as upside, be suspicious.

What does zero-knowledge look like in a notes app?

A zero-knowledge notes app derives or holds its key on your device, encrypts every note and attachment before sync, unlocks with biometrics by releasing a stored key rather than skipping encryption, and gives you a recovery key instead of an email reset. Standard Notes and Notesnook are examples of this design.

1. Client-side keys. Password or device key, then encryption on the device. The server sees nothing readable.
2. Full scope. Note bodies, titles, attachments and ideally metadata are encrypted, not just the text.
3. Biometric unwrap, not bypass. Face ID should release a key held by the Secure Enclave or Keychain, not replace the encryption.
4. User-held recovery. A recovery key or phrase you store offline. If support can restore your data without it, it is not zero-knowledge.

Red flags on a privacy page: “encrypted on our servers,” support-assisted recovery, no named cipher, no word on scope, and no code or audit to check. Our comparison of encrypted notes apps applies these tests app by app, and the head-to-heads with Standard Notes and Notesnook go deeper.

Is Secure Notes zero-knowledge?

No, and we would rather say so than stretch the term. Secure Notes avoids the server problem differently: PixelPort runs no server that stores your notes, and you do not create an account. On iPhone, notes stay on the device under iOS Data Protection and the app's passcode and Face ID lock. Optional sync (Pro) goes to your own private iCloud.

Those iCloud copies get Apple's standard iCloud protection, and become end-to-end encrypted only if you turn on Advanced Data Protection. The iPhone app's Pro Encrypted Backup file (version 3.8.1 and later) is encrypted on your device with AES-256-GCM using a password you choose, so only you can open it. On Android, the whole vault is AES-256 encrypted on the phone (SQLCipher, key protected by your PIN via Argon2id) and is never uploaded anywhere.

The 12 recovery words in Secure Notes get you back into the app if you forget your passcode; they are not an encryption key. See recovery phrases for notes apps for the difference, and our security page for the full model.

Frequently asked: zero-knowledge encryption