What end-to-end encryption actually means (and what it does not)
A plain explanation of end-to-end encryption: what it protects, how it differs from "encrypted in transit", and the questions worth asking before you trust a product that claims it.
Almost every service you use claims to encrypt your data. Very few of them mean the same thing by it, and the difference decides whether a company can read your messages.
This is the plain version, without the marketing.
Three different things people call "encrypted"
Encrypted in transit. Your data is scrambled while it travels between your device and a company's servers. This is what the padlock in your browser means. It stops somebody on the same café wifi from reading your traffic. It does not stop the company at the other end from reading it, because they hold the key — they have to, in order to receive it.
Essentially every website does this now, and a product advertising it as a privacy feature is telling you nothing.
Encrypted at rest. Data sitting on a company's disks is stored scrambled. This protects against someone stealing a hard drive out of a data centre, and against a certain class of backup leak. It does not stop the company reading your data, because their systems decrypt it routinely in order to work.
End-to-end encrypted. Data is encrypted on your device, before it is sent, with a key the service never has. It stays encrypted the entire way, and is only decrypted on the recipient's device. The company in the middle is carrying a sealed envelope it cannot open.
Only the third one changes what the service itself can see. The first two protect you from outsiders; the third protects you from the service.
What "the service never has the key" really requires
The claim is easy to make and harder to build, because a lot of ordinary product features quietly require the ability to read your data:
- Password reset by email. If a company can restore your access after you forget a password, they can decrypt your data, so they must have a key. Some services solve this carefully with key escrow you control; many just have the key.
- Server-side search. Searching your messages on the server means the server can read them.
- Previews and thumbnails generated in the cloud. Generating an image preview requires seeing the image.
- Spam and content filtering on content. Scanning message contents requires reading message contents.
- "View in browser" links that work without you being logged in.
None of these are automatically sinister — they are all genuinely useful. But each one is incompatible with the strong version of end-to-end encryption, and a product that offers them alongside an end-to-end claim is doing something more complicated than the marketing suggests.
The tell is simple: ask what happens if you lose your key or password. If the answer is "we can restore your access", the service can read your data. If the answer is "your data is unrecoverable, including to us", the claim is at least structurally possible.
How it works, roughly
The mechanism most modern systems use is public-key cryptography, and the useful mental model is a padlock.
You generate two matching pieces: a public key and a private key. The public key is like an open padlock you hand out freely — anyone can use it to lock a box addressed to you. The private key is the only thing that opens those padlocks, and it never leaves your possession.
So when someone wants to send you something:
- They get your public key. It is public; it does not matter who sees it.
- Their device encrypts the message with it.
- The result is unreadable to everyone, including whatever server carries it.
- Your device decrypts it with your private key.
In practice, systems combine this with faster symmetric encryption — the public key encrypts a one-off random key, and that key encrypts the actual content, because public-key operations are slow on large data. The security property is the same.
The important structural consequence: the server's role is reduced to delivery. It routes sealed envelopes. It cannot inspect them, and neither can anyone who compels, hacks, or buys it.
What end-to-end encryption does not protect
This is the part most explanations skip, and it matters more than the mechanism.
It does not protect a compromised device. Encryption happens on your computer or phone. If malware, a shared account, or another person has access to that device, they have access to your messages after decryption. No amount of cryptography helps once the endpoint is the problem — the "ends" in end-to-end are the weak points.
It does not control your recipient. Once someone can read what you sent, they can screenshot it, copy it, forward it, or keep it forever. Encryption governs who can read a message in transit, not what a reader does next. "Disappearing messages" are a courtesy, not an enforcement mechanism.
It does not hide that you communicated. The content is opaque, but metadata often is not: that a connection happened, when, how large it was, and sometimes who with. Different systems leak different amounts of this. Anyone whose threat model includes "someone must not know I contacted this person at all" needs to think about metadata specifically, not just encryption.
It does not protect against the code you are running. If encryption happens in a web page, you are trusting that the page served to you today does what it did yesterday. This is a real limitation of browser-based tools compared to signed, audited native apps — worth knowing, and worth weighing against the fact that browser tools require no installation and no account.
It does not survive lost keys. If the service cannot read your data, the service cannot recover it either. That is the same property viewed from the other side, and it means backup is your responsibility.
Questions worth asking
Before trusting any product's encryption claim:
- What happens if I lose my password or key? "We can restore it" means they hold a key.
- What can you see? A good answer is specific — public keys, message sizes, timestamps — rather than "we take your privacy seriously".
- What can you hand over if compelled? A service that can answer this precisely has thought about it.
- What algorithms? Named, standard primitives, not "military-grade encryption", which is not a thing.
- What is not protected? A product that will not tell you its limits is describing a wish, not a design.
Where AliasCloak sits
AliasCloak takes the strict version, with the consequences that follow.
Your keys are generated in your browser and the private key never leaves it. We hold your public key and a readable alias so people can reach you. Messages and files arrive as ciphertext we have no key for, and are deleted once your device confirms it decrypted them.
That means there is no password reset, no server-side search, and no way for us to recover a mailbox whose private key is gone — including under a court order, because we cannot produce something we do not have. It also means you should back up your private key before you rely on it for anything.
And the limits above apply to us exactly as they apply to everyone else. They are written out on the security page, including the parts that are not flattering.
Keep reading
Private by design
