Encryption
Zero-knowledge, explained
The technical meaning is precise. A zero-knowledge proof lets one party demonstrate knowledge of something — a password, a key, a valid transaction — without transmitting it. The construction is genuinely remarkable and underpins real systems in authentication and in privacy-preserving ledgers.
The marketing meaning is unrelated to any of that. When a storage or messaging vendor says "zero-knowledge architecture", what is nearly always meant is client-side encryption with keys the server never receives — which is a real and valuable property, but its actual name is end-to-end encryption. Borrowing the cryptographic term adds no information and makes the claim harder rather than easier to verify.
How Enclessa uses it
Enclessa does not describe itself as zero-knowledge, and will not. The accurate statement is narrower and more useful: direct messages and group direct messages are end-to-end encrypted with MLS and the server holds no key for them; channels are, by default, managed rooms the server can read; and metadata, account data and billing data are visible to Enclessa in every case. Anywhere the guarantee is weaker, Enclessa names the room mode instead of reaching for a stronger word.