Glossary

ML-KEM

ML-KEM is a post-quantum method standardized by NIST that lets two devices agree on a shared secret key without sending that key over the network.

What is ML-KEM?

ML-KEM stands for Module-Lattice-Based Key-Encapsulation Mechanism, a method built on the mathematics of lattices. NIST published it in 2024 as the standard FIPS 203, based on the earlier CRYSTALS-Kyber design. According to current knowledge, it is designed to resist attacks even from quantum computers.

How ML-KEM works

ML-KEM does not encrypt messages itself; instead, it gives two parties a shared secret without transmitting that secret openly. One side publishes a public key, and the other uses it to “encapsulate” a random secret that only the holder of the matching private key can unwrap. That secret is then used to derive a key for a fast symmetric cipher such as AES. The standard defines three levels: ML-KEM-512, ML-KEM-768 and ML-KEM-1024.

At Friendstime

Friendstime uses ML-KEM-768, the middle of the three levels, for the key exchange for the text of private chats. If an account does not have an ML-KEM key yet, RSA-OAEP is used instead; a combination of both methods (hybrid) does not currently exist. The chat text is end-to-end encrypted, and the server only stores the encrypted text. Public content such as posts and profiles, by contrast, is not end-to-end encrypted.

Frequently asked questions

Is ML-KEM the same as Kyber?

ML-KEM is based on CRYSTALS-Kyber, but some details were changed for the FIPS 203 standard, so the two are not fully interchangeable.

What does the number in ML-KEM-768 mean?

It names the parameter set. ML-KEM-768 is the middle of three levels and is assigned to NIST security category 3.

Related terms

Related topics

Try Friendstime

Signing up takes a minute and is free.

All terms in the glossary →