Zero-Knowledge Cloud Storage: What It Means for Your Memories
Zero-knowledge cloud storage keeps your files unreadable to the provider. Learn how it works, which metadata stays visible, and the honest trade-offs to weigh.
Key takeaways
- Zero-knowledge cloud storage means the provider stores your files but never holds the keys needed to read them.
- The term describes file contents, not everything: account details, file sizes, file types and timestamps usually remain visible.
- The honest trade-off is that a lost password and recovery key can mean lost files, because nobody else can open them.
- Server-side conveniences such as face search and server-built thumbnails are limited, because the server cannot see inside your photos.
- A trustworthy zero-knowledge service lets you export everything and open that export without the service.
Zero-knowledge cloud storage is a service that stores your files without ever holding the keys needed to read them. Your files are encrypted on your own device before upload, so the provider keeps only scrambled data, and only you, plus anyone you choose to share with, can turn it back into photos, letters and recordings. The trade-off is real: if you lose your keys, the provider cannot recover your files for you.
For family memories, that combination of strong privacy and personal responsibility is worth understanding before you move an archive. Here is what zero-knowledge means, what it hides, what it still reveals, and how to use it without locking yourself out.
What zero-knowledge cloud storage means
The phrase borrows from cryptography, where a "zero-knowledge proof" is a specific technique for proving that a statement is true while revealing nothing beyond that fact.1 Storage companies use the term more loosely. When a provider calls its storage zero-knowledge, it means the provider has no knowledge of your file contents, because it never has your key.
A storage unit is a helpful picture. You rent the unit, bring your own padlock, and keep the only key on your keyring. The facility knows your name, your unit number, how big the unit is, and when you come and go. It has no idea whether the boxes inside hold wedding albums or winter coats, and it cannot open the door to check.
One phrase is easy to confuse with it. "Zero trust" is a model for securing an organization's own network, set out by NIST: no user or device is trusted because of where it connects from, and every request is verified again.2 It describes how a company guards its systems, not whether it can read what you store, and many zero-trust deployments inspect traffic as part of the design. Zero-knowledge is the term to look for when the question is who holds the keys.
In practice, zero-knowledge storage relies on client-side, end-to-end encryption. If you want the full mechanics, our guide to end-to-end encryption explains how keys move between people. The short version follows.
How zero-knowledge storage works behind the scenes
Most zero-knowledge services follow a similar pattern:
- You choose a password or passphrase for your encrypted space.
- Your device runs it through a memory-hard key derivation function, often Argon2id, which makes every password guess slow and expensive for an attacker.3
- The result protects a random master key that is generated on your device.
- Each file gets its own random key, and the file is encrypted with a cipher such as AES-256-GCM before it leaves your device.45
- The file keys are wrapped with your master key, and only encrypted files and wrapped keys are stored on the provider's servers.
Careful services encrypt file and folder names too, so a server sees anonymous objects rather than names like "Dad letters 1971." Argon2id itself is a public, well-documented design, described in RFC 9106 (opens in a new tab).
The provider still runs the servers, handles your sign-in and keeps the encrypted data safe from loss. What it does not have is any step in which your readable files or your unwrapped keys pass through its hands.
Zero-knowledge vs standard cloud storage
Standard cloud storage also encrypts your files, usually in transit and on the provider's disks. The difference is who holds the keys. Side by side:
- Who can read your files: with standard storage, the provider can when its systems need to; with zero-knowledge storage, only you and the people you share with.
- Password reset: standard storage can restore full access after an email reset; zero-knowledge storage needs your vault password, recovery key or another method you prepared.
- Search and previews: standard storage can analyze photos on its servers; zero-knowledge storage can only do this on your device, if at all.
- A breach at the provider: with standard storage, readable files may be exposed if keys are taken too; with zero-knowledge storage, stolen files stay scrambled without your key.
- Metadata: both can usually see account details, file sizes and upload times.
Neither model is careless. They suit different material, and plenty of families use one for everyday snapshots and the other for the archive.
What zero-knowledge does not hide: the metadata
Zero-knowledge describes your file contents. It does not describe everything a service knows about you, and an honest provider will say so. Think of a letter in the mail: the post office cannot read the letter, but it can read the envelope.6
Metadata a zero-knowledge service may still see includes:
- Your email address, name and billing details.
- When you sign in, from which devices, and network details such as IP addresses in its logs.
- How many files you store, their sizes and their file types.
- When files were uploaded or changed.
- Who you have named as contacts or recipients, and any delivery schedule you set.
- The shape of your sharing, such as how many people can open a given folder.
Some of this is unavoidable. A service cannot enforce a storage limit without knowing file sizes, and it cannot notify your sister without knowing her email address. Apple, for instance, keeps modification dates and checksums of photo data under its standard protection even when Advanced Data Protection is on, so it can sort and de-duplicate storage without access to the photos themselves.7 The useful question is not whether metadata exists but which metadata, and whether the provider tells you plainly.
Photos also carry details inside the file itself. When a phone's camera has location access, those details include the coordinates where each picture was taken.8 In a zero-knowledge vault, those details are encrypted along with the photo. If you share the original file, the person you send it to may be able to read them.8
The honest trade-offs of zero-knowledge storage
Zero-knowledge storage is a deliberate trade. You gain privacy from the provider and give up some things the provider could otherwise do for you.
A lost key can mean lost files
This is the big one. If you forget your vault password and have no recovery key or other prepared method, the provider cannot reset your way back in. There is no back door, because a back door would mean the provider could read your files. Picture a time capsule whose only key went into the lake along with a phone. Recovery keys, passkeys and similar options exist precisely so that never happens, but you have to set them up in advance.
No server-side search, faces or thumbnails
A conventional photo service can recognize your grandmother's face across ten thousand pictures, find every photo of a birthday cake, and generate previews on its servers. A zero-knowledge service cannot see the photos, so it cannot do any of that on its side. Some services rebuild a few conveniences on your device after you open the vault, which can be slower on a large library. Many families find that a fair price for the archive that matters most.
Sharing needs keys, not just links
Sharing a file means giving another person a way to decrypt it, either through their own account keys or a key you hand over separately. It takes a minute longer than pasting a public link, and it means the person you share with really is the only one who can open it.
Some features need your consent to see plaintext
Features that read your content, such as AI writing help or transcription, need readable data. In a zero-knowledge design, these should be optional, clearly explained, and switched on only when you consent to send specific material.
Your devices matter more
Your files are decrypted on your phone or computer, so keeping those devices updated and locked matters. Encryption protects stored copies; it cannot protect a screen someone else is reading.
Who zero-knowledge cloud storage is right for
Zero-knowledge storage suits material that is private, irreplaceable and meant to last. Think of scanned adoption papers, a parent's letters about an illness, a decade of journals, recordings of a grandfather singing in Portuguese at the kitchen table, or the documents your children will one day need.
It is less necessary for the thirty photos of a parking garage you took to remember where you left the car. Many families settle on two homes: a convenient everyday photo service for the daily stream, and a zero-knowledge vault for the archive. Either way, no single service should hold your only copy, which is where the 3-2-1 backup rule helps.
How to set up zero-knowledge storage without losing access
A few habits make the trade-offs comfortable:
- Choose a long passphrase for the vault that you do not use anywhere else.
- Save the recovery key the moment it is offered, on paper, in two separate safe places.
- Add a passkey or a second recovery method if the service supports one.
- Sign out and back in once, then open a file, to prove you can.
- Make an export and confirm you can open it.
- Decide who in your family should be able to reach the vault one day, and how.
That last step is easy to postpone. Our guide on how to pass down passwords covers calm, practical ways to plan it.
Questions to ask a zero-knowledge provider
Before you move a family archive, ask:
- If I reset my sign-in password, can I still see my files? In a zero-knowledge design, a reset alone should not open the vault.
- Which metadata can you see, and is it listed somewhere I can read?
- Are file names encrypted as well as file contents?
- Can I export everything, and can I open that export without your service?
- Where are the encrypted files stored, and with how many providers?
- Has the design been independently reviewed, and if not yet, do you say so?
That last question matters: when cryptographers at ETH Zurich examined five end-to-end encrypted cloud storage services with more than 22 million users between them, they found severe cryptographic weaknesses in four, including ways a malicious server could inject or tamper with files.9
If you are weighing several services, our checklist on whether cloud storage is safe for family photos goes further.
How Henedo approaches zero-knowledge storage
Vault content is encrypted in your browser with AES-256-GCM before upload, and file names are encrypted too. A separate vault password, run through Argon2id, protects your master key, so signing in to your account is not the same as opening your vault. Henedo cannot recover a lost vault password or key, which is why the app asks you to save an offline recovery key and supports passkeys.
The limits are part of the design. Account, storage and delivery metadata, including file sizes, types and timestamps, remains visible to the service. Optional AI composition sends selected written answers only after you consent. Encrypted files are currently held with one storage provider, so the owner export includes every encrypted file plus an offline reader that opens them without the service, as described on the portability page. The full key hierarchy is on the security page.
Sources
Each number in the article links to the published work it draws on.