How end-to-end encryption works
End-to-end encryption uses public-key cryptography. Each person has a pair of keys: a public key they can share with anyone, and a private key that never leaves their device.1
- Keys are made on your device. Your private key is created where you are and stays under your control.
- The sender locks the message. Their app fetches your public key and uses it to encrypt the message before it leaves their device.
- The service only relays. Servers pass along and store only ciphertext, without the keys to open it.
- You unlock it. Only your private key can decrypt the message, on your device.
In practice, the public key protects a fresh random key, and that key encrypts the message with a fast cipher. That combination is called hybrid encryption.
End-to-end encryption vs other kinds
| Approach | Protected on the network? | Can the service read it? |
|---|---|---|
| Encryption in transit (TLS) | Yes | Yes, once it arrives |
| Encryption at rest, provider keys | Depends | Yes |
| End-to-end encryption | Yes | No |
Why it matters
If a service can read your messages, so can anyone who hacks it, any employee with the right access, and anyone who sends it a legal demand. End-to-end encryption removes the service from that list. A data breach exposes only ciphertext.
It is worth knowing what it covers. End-to-end encryption protects the content of a message. Depending on the system, details like who you talked to and when may still be visible to the service. And it protects messages in transit and in storage, not on a device that is already compromised.
Where you’ll see it
- Messaging apps, many of which now use it by default. Signal’s protocol, for example, now adds a post-quantum key exchange to its end-to-end encryption.3
- Email, through standards like OpenPGP and S/MIME,45 and through encrypted email services that handle the keys for you.
- Cloud storage and backups, where some services offer it as an option.
A newer concern is the quantum computer. Most end-to-end systems still rely on elliptic-curve keys that Shor’s algorithm could one day break, so messages recorded today could be read later. Post-quantum key exchange, such as ML-KEM, closes that gap.
Related terms
Sources
- Knodel, Celi, Kolkman and Grover, IETF, Definition of End-to-end Encryption (Internet-Draft).
- Signal, The X3DH Key Agreement Protocol, section on authentication.
- Signal, The PQXDH Key Agreement Protocol.
- IETF, RFC 9580: OpenPGP (July 2024).
- IETF, RFC 8551: S/MIME Version 4.0 Message Specification (April 2019).
Questions about end-to-end encryption
What does end-to-end encryption mean?
That a message is encrypted on the sender’s device and can only be decrypted on the recipient’s. Nobody in between, including the service, can read it.
Is end-to-end encryption safe?
It is the strongest protection widely available for messages. Its main limits are outside the encryption itself: a hacked phone, or details like who you contacted, which some services can still see.
Can the police read end-to-end encrypted messages?
Not by asking the service, because the service doesn’t have the keys. They would need access to a device at one end of the conversation.
What is the difference between end-to-end encryption and TLS?
TLS protects data between your device and a server, which then decrypts it. End-to-end encryption keeps data locked all the way to the recipient, so the server never sees it in readable form.