NIP-4E is an open draft, proposed by fiatjaf, for sharing private data between a user’s own devices without every device holding that user’s main Nostr identity key. It is not merged and remains a draft/optional proposal.

The Problem It Addresses

Many existing NIPs, including NIP-51 lists and NIP-60 Cashu wallets, encrypt data from a user to themselves using the identity key so it can be read back later on any device. That breaks down when the identity key isn’t directly accessible, for example when a remote signer is protected by FROST threshold shares, MuSig2, or a hosted secure enclave, since encrypting and decrypting then requires a round trip to that signer every time. It also makes offline encryption impossible whenever the signing key lives in a remote bunker.

How It Works

NIP-4E separates a per-device “client key” from a shared “encryption key” that is not the user’s identity key:

  1. The first client a user sets up generates a random encryption keypair and announces its public half in a kind:10044 event signed by the user’s identity key.
  2. Any other client that wants to encrypt or decrypt data for that user computes its Diffie-Hellman shared secret against the announced encryption key rather than the identity key.
  3. When a second device installs a new client, that client generates its own local “client key” and publishes a kind:4454 announcement (also signed by the user’s identity key) asking the first client to share the encryption key with it.
  4. The original client detects the new kind:4454 announcement, encrypts the shared encryption key to the new client’s key using NIP-44, and publishes it so the new client can decrypt and use it going forward.

The result is that encryption and decryption never require asking the identity-key signer at all once a client holds the shared encryption key locally, and a remote-signer setup (FROST, MuSig2, hosted enclave) can be used for identity while ordinary encryption stays fast and works offline.

Why It Matters

NIP-4E is cited as the foundation for other proposals that need a drive-wide or account-wide symmetric key without depending on a remote signer for every encrypt/decrypt call, including a private encrypted drive proposal (PR #2412) and a narrower NIP-17-specific version of the same idea (PR #2361). Both remain open alongside NIP-4E itself, making this an active, unsettled area of the protocol rather than a finished building block.

Primary sources:

Mentioned in:

See also: