Cybersecurity

What Is End-to-End Encryption?

End-to-end encryption, usually shortened to E2EE, is the strongest form of message encryption: only the sender and the recipient can read the messages — not the company running the app, not the internet provider, and not anyone who intercepts them in transit. This guide explains how it works, what it protects, and the important things it does not protect.

By Abdul Hannan · Published 14 November 2026 · Updated 14 November 2026

Note. Encrypted does not mean private. End-to-end encryption hides the content of your messages, but it does not hide the metadata: who you messaged, when, how often, and how much. That distinction matters, and it is the core of this guide.

What end-to-end encryption means

In end-to-end encryption, the messages are encrypted on the sender's device and decrypted only on the recipient's device. The keys are held by the people talking, not by the service in the middle. The company's servers carry the scrambled messages from one end to the other, but the company has no key to read them — and a court order, a breach, or a rogue employee at the company cannot make the content appear, because the company does not hold what would unlock it.

Compare this with ordinary transport encryption. When you send a message over a secure connection, the connection itself is encrypted, so no one watching the network can read it. But the message is decrypted when it arrives at the company's server, stored in readable form, and re-encrypted for the journey to the recipient. The company can read it — for spam filtering, for search, for adverts, or because it is legally required to. End-to-end encryption removes that middle step: the server never sees the readable message at all.

The most familiar way to recognise E2EE is a security code or safety number: a number or QR code that two people can compare to confirm that no one is sitting in the middle of their conversation. If the codes match, the channel is genuinely between them.

How it works, without the maths

Each device generates a key pair: a public key, which can be shared openly, and a private key, which never leaves the device. When you start a conversation, the two devices exchange public keys — through the server, which can see the public keys but cannot use them to read the messages — and agree a shared secret that only they know. Messages are encrypted with that shared secret before they leave the sender's phone, and decrypted only when they arrive on the recipient's phone.

The server's role is delivery, not reading. It stores scrambled messages until the recipient is online, and it can route them, count them and time them — but it cannot decrypt them. If the server is breached, the attacker gets ciphertext. If the company is ordered to hand over the messages, it can hand over only what it holds — the scrambled version.

Group chats work the same way, with the shared secret extended to every member of the group. Add a member, and the keys are updated; remove one, and they lose access to new messages. The details vary between apps, but the principle is the same: the keys live at the ends, not in the middle.

What end-to-end encryption protects

It protects the content of your messages from everyone except the other person in the conversation: the app company, your internet provider, anyone on a public Wi-Fi network, and anyone who breaches the company's servers. That protection matters most for sensitive conversations — health, finance, journalism, legal advice, and any discussion you would not want exposed in a breach.

It also protects a conversation's history. Because the messages are stored encrypted on the company's servers, a breach of those servers does not expose years of your conversations. The protection is only as strong as the devices at the ends, though: if your phone is unlocked and opened by someone else, the messages are readable there, because that is where they are decrypted.

It does not protect you from the person you are talking to. Any message you send can be screenshotted, copied, forwarded or stored by the recipient, and no encryption can prevent that. Disappearing messages help, because they remove the copy from both devices after a set time — but a screenshot taken before they disappear still exists.

What it does NOT protect: metadata

This is the part most explanations skip. End-to-end encryption hides what you said, but it does not hide who said it, to whom, when, how often, or how large the message was. That surrounding information is called metadata, and it is visible to the service, because the service needs it to deliver the messages. It can also be visible to anyone watching the network, and it can be collected, stored and, in some cases, handed over under a legal order.

Metadata can reveal a great deal on its own. A record showing that you messaged a particular doctor's office every week, at the same time, tells a story even without the words. A pattern of late-night messages to a journalist, a lawyer, or a competitor tells another. That is why privacy is not the same as encryption, and why a messaging app can be genuinely encrypted and still collect a significant record of your life.

Backups are the other gap. Many apps encrypt messages in transit but store backups in the cloud unencrypted, or encrypted with a key the company holds. When you turn on cloud backup for an encrypted chat app, check what the backup actually protects — a backups section in the app's settings will say, and the honest answer is sometimes less reassuring than the headline about encryption.

Encryption hides what you said — not who, when, or how often.
Metadata is visible to the service because it must deliver messages.
Cloud backups may not be end-to-end encrypted — check the settings.
The recipient can always screenshot, copy or forward a message.

How to tell if a chat is end-to-end encrypted

Check the app's own description of the conversation: most encrypted apps show a notice — a lock icon, a message saying messages are end-to-end encrypted, or a security code you can compare with the other person. If an app does not say a conversation is end-to-end encrypted, assume it is not. Group chats and backups may have different settings from one-to-one chats, so check each one.

Two practical checks matter. First, verify the security code with the person you talk to most — compare the number or scan the QR code, in person or through a channel you trust. If the code changes unexpectedly, that can mean the other person changed devices — or that someone has inserted themselves in the middle. Second, check your backup settings: turn on encrypted backup where it is offered, or turn cloud backup off for the app if you would rather keep the messages only on your devices.

Be wary of a claim that an app is encrypted without detail. End-to-end encryption is a specific, checkable property, not a marketing word — the apps that genuinely offer it explain it, name the protocol, and let you verify the security code.

Choosing a messaging app

If encryption matters to you, choose an app that offers end-to-end encryption by default for one-to-one chats, and check whether groups and backups are covered too. Look for an open, published protocol that independent experts can review, and for a company that publishes transparency information about the requests it receives for data.

Default settings matter more than features. An app that encrypts by default protects everyone who uses it, including people who never open the settings. An app that offers encryption as an option protects only the people who find the switch. And remember that encryption is one layer: a strong device lock, an up-to-date app, and care about what you share are the layers that sit on top of it.

Sources

National Cyber Security Centre (UK) — Encryption guidance — last checked 14 November 2026.
NIST — Cryptographic standards — last checked 14 November 2026.
CISA — Cybersecurity best practices — last checked 14 November 2026.