what does zero knowledge mean zero knowledge meaning zero-knowledge proof zero-knowledge encryption zero knowledge vs end to end encryption encryption

What does zero-knowledge mean? Both meanings, explained

Adrian Maverick · · 10 min read

Search "zero-knowledge" and you get two sets of answers that seem to disagree. One says it is a way to prove something is true without revealing why. The other says it is a way to store data so the company holding it cannot read it. Both are right. They are two different ideas sharing one name.

Zero-knowledge means one of two things. In cryptography, a zero-knowledge proof lets you prove a statement is true without revealing anything beyond the fact that it is true. In privacy products, zero-knowledge encryption means a service stores your data encrypted with keys only you hold, so the provider cannot read it.

What does zero-knowledge mean? The two senses side by side

Zero-knowledge proof Zero-knowledge encryption
What it is A protocol between a prover and a verifier A product design for storing data
Question it answers "Can I convince you I know X without telling you X?" "Can you hold my data without being able to read it?"
Where you meet it Cryptocurrency, "zk-rollups," digital ID, age checks Password managers, cloud storage, encrypted email
Where it came from A 1985 paper by Goldwasser, Micali, and Rackoff Borrowed later by privacy products
Who "knows nothing" The verifier learns nothing but the truth of the claim The provider learns nothing about your content

Wikipedia's disambiguation page for "zero knowledge" lists the proof, two variants of it, and "zero-knowledge service," the privacy-product sense. The blur is so common that even vendors mix them. Bitwarden's own explainer opens by defining zero-knowledge encryption as "a cryptographic technique that allows one party to prove knowledge of a secret to another party without actually revealing the secret itself," which is the proof definition, and then describes a model "where even the provider has no access to your plaintext," which is the storage one.

Zero-knowledge proofs: convincing someone without telling them anything

The idea came from a 1985 paper by Shafi Goldwasser, Silvio Micali, and Charles Rackoff, "The Knowledge Complexity of Interactive Proof-Systems." It was foundational enough that the three shared the first Gödel Prize in 1993 with László Babai and Shlomo Moran, who developed interactive proofs independently.

A zero-knowledge proof example: the cave

Picture a ring-shaped cave with a locked door splitting the far end into two paths. You claim you know the password to the door. I want to be convinced without learning it. I wait at the entrance, you walk in and take a path at random, and then I shout which side I want you to come out of.

If you know the password, you can always come out the side I name, because you can open the door and switch. If you are bluffing, you have a fifty-fifty chance each time. Do it twenty times in a row and the odds you bluffed every round are about one in a million. I end up convinced, and I still have no idea what the password is.

The three properties

Formally, a zero-knowledge proof must be:

  • Complete. If the statement is true, an honest prover can convince an honest verifier.
  • Sound. If the statement is false, a cheating prover cannot convince the verifier, except with tiny probability.
  • Zero-knowledge. If the statement is true, the verifier learns nothing beyond that fact.

Where you meet zero-knowledge proofs

Mostly in cryptocurrency and digital identity. Blockchains use them ("zk-SNARKs," "zk-rollups") to show a batch of transactions is valid without replaying every detail. Newer digital ID designs use them to prove you are over 18 without revealing your birth date. If a page mentions blockchains, rollups, or "proving without revealing," it means this sense.

Zero-knowledge encryption: a provider that cannot read your data

This is the sense most people searching the term want. When a password manager, backup service, or email provider says "zero-knowledge," it is making a claim about who holds the keys.

Ordinary cloud services encrypt your data but keep the keys. Your files are scrambled on their disks and unscrambled on their servers whenever the company, or anyone who breaks in, needs them. The lock is real, but the key sits in the same building.

Zero-knowledge encryption removes the provider's usable key. Data is encrypted and decrypted on your device with a key derived from a secret the provider never sees, usually your master password run through a slow key-derivation function. Bitwarden, for example, says your master password "is never stored or accessed by the provider." The company stores ciphertext it cannot open.

Cryptographers will point out that "zero-knowledge encryption" is a borrowed marketing term. The precise description is client-side encryption with keys that never leave your control in usable form. Different companies use different names for that design. In email you will often see zero-access encryption, which is what we call it at Secria: stored mail is sealed with keys the provider never holds.

Zero-knowledge vs end-to-end encryption

End-to-end encryption describes a message in transit: only the sender and recipient can read it, and the servers in between pass along ciphertext. Zero-knowledge (or zero-access) describes data at rest: what the provider stores, it cannot open.

A good private service usually does both. You can have one without the other. A chat app can be end-to-end encrypted in transit but keep readable backups in a cloud whose keys it controls. And a storage service can be zero-knowledge for your files without sending anything to anyone.

Zero-knowledge vs zero trust

These get confused because both start with "zero." Zero trust is a network security approach for organizations: never assume a device or user is safe just because it is inside the network, and verify every request. It has nothing to do with who holds encryption keys.

How to test a "zero-knowledge" claim

"Zero-knowledge" is not a certified label. Anyone can print it on a website. Five questions sort real designs from marketing:

  1. Where are the keys created and kept? The answer you want: generated on your device, and never on the provider's servers in a usable form. An encrypted key blob the provider cannot open is fine; a key it can use is not.
  2. What happens if you forget your password? If the provider can reset it and hand your data straight back, it had access. A real design depends on something only you hold, such as a recovery phrase or recovery key.
  3. Is there a published design or independent review? A whitepaper that explains the cryptography, open-source client code, or a published audit is how a provider shows the architecture matches the marketing.
  4. What does the provider say it can hand over under a court order? Honest providers say: encrypted data plus limited account metadata, because that is all they hold.
  5. What does the provider say about metadata? Content encryption covers the letter, not always the envelope. A provider that tells you plainly what it can see is easier to trust. (More on this in our guide to metadata privacy.)

How Secria answers those questions

Here is how Secria Mail answers them. Your keys are generated on your device, and your private key is encrypted there before it is stored, so your mail sits under zero-access encryption, sealed with keys we never hold. We cannot reset your password and restore your mailbox; the 12-word recovery phrase you keep is what gets you back in. The design is written up in a public whitepaper. Encryption is hybrid post-quantum (ML-KEM-1024 with X25519) on every plan, including free. And our FAQ states what we can produce under a legal request: not your stored mail, because we do not hold the keys to it.

If you are choosing an email provider specifically, our older post on what zero-knowledge email actually means covers the email-specific checklist.

What zero-knowledge does not do

  • It does not cover what was never encrypted. Metadata like account details and timestamps usually sits outside the content encryption.
  • It depends on the app you run. Decryption happens in software the provider ships, which is why published designs, open code, and audits matter.
  • Recovery is your job. Without a provider backdoor, there is no provider rescue. Store your recovery phrase or key somewhere safe and offline.

There is also a longer clock. Most encryption in use today relies on math a large quantum computer could break, so data recorded now could be opened later. That is why NIST finalized its first post-quantum standards in 2024. Here is post-quantum encryption explained without the jargon.

Frequently asked questions

Is zero-knowledge the same as end-to-end encryption?

No. End-to-end encryption protects a message between sender and recipient. Zero-knowledge describes stored data the provider cannot decrypt. Strong privacy services usually combine both.

Does zero-knowledge mean the company knows nothing about me?

No. It means it cannot read your content. Most services still see some account information, such as your email address and when you sign in. Check each provider's privacy policy for specifics.

Is a zero-knowledge proof the same as zero-knowledge encryption?

No. A zero-knowledge proof is a protocol for proving a statement without revealing why it is true, used heavily in crypto and digital ID. Zero-knowledge encryption is a storage design where the provider cannot decrypt your data. Same phrase, different fields.

What is zero-access encryption?

It is the term many email providers use for the same goal: stored mail is encrypted with keys the provider does not hold, so it cannot read or hand over the contents.

Both meanings answer the same instinct from different directions: reveal as little as possible while still getting the job done. Now you can tell which one a page is talking about, and whether the claim holds up.

Sources

Secria fact-checks every post against primary sources. Spotted something wrong or out of date? Email hq@secria.me and we will correct it.

Back to all posts