Resources
BetterSign treats key creation, rotation, revocation and recovery as first-class signed lifecycle events. Peers can discover, verify and apply them without the stable identity changing.
Resources › Key Lifecycle
The protocol. This page describes the BetterSign protocol and its open-source core. For what the BetterSign app and products do today, see Products.
A complete key-management system needs creation, rotation, revocation and recovery. BetterSign models all four as provenance-log state transitions, not one-off file distribution events.
Lifecycle events are first-class citizens in BetterSign. They are not side-channel notifications, support tickets or emergency ceremonies. They are signed entries in the same verifiable history that defines current identity state.
The identity remains the VLAD. Keys are current state under the VLAD plog, so a follower can verify lifecycle history and derive the present trusted key set.
Key creation generates initial cryptographic material, assigns intended operations, and publishes the resulting public state to the BetterSign network.
BetterSign does not require every deployment to store secrets the same way. The generated public keys and policy become verifiable plog state. The private-key storage strategy stays the operator’s choice.
Private keys remain under the chosen local custody model.
| VLAD | Stable identity address |
/keys/primary | Ordinary operational signing authority |
/keys/recovery | Higher-priority recovery and administrative authority |
/keys/encrypt | Public encryption or KEM material for sealed messages |
| policy | Lock scripts and proofs required for future updates |
Rotation invalidates old key material and replaces it with new material while preserving the identity. It can be triggered by expiration, suspected compromise, cryptographic agility, or a change in policy.
A normal rotation is authorized by the current accepted key or policy. Followers do not need a manual out-of-band key file. They verify the plog update and derive the new current state.
Key changes should be expected. BetterSign makes rotation a routine state transition, so operators can shorten key lifetimes, adopt stronger algorithms, and recover from exposure without treating every change as an outage.
BetterSign can let a peer jump to the latest advertised head while still validating the intermediate chain, so long histories stay practical without giving up lifecycle authentication.
See rotation and propagation in action: one signed entry, and every follower verifies and switches on its own.
Revocation invalidates a key without immediately naming a replacement. It uses the same publication and verification path as rotation, but the resulting state marks the key as no longer usable.
Followers that watch the relevant plog paths can react by removing SSH keys, refusing a signing authority, updating service registries, or forcing local policy to reject the revoked material.
Recovery is the highest-risk lifecycle operation, because it restores authority after loss or compromise. BetterSign handles recovery through explicit higher-priority authority, rather than asking a third party to reset the identity.
A common model is P0/P1 authority. A P1 primary key handles day-to-day updates. A better-protected P0 recovery key can override or replace the P1 key when compromise or loss occurs.
This maps directly to the standard recovery-before-primary lock behavior: a valid recovery proof can authorize recovery-grade updates before the normal primary proof is considered.
Outcome. P0 recovery wins contested updates. P1 primary handles normal operations.
BetterSign does not require one custody model for all deployments. Operators can choose key storage that matches their threat model while publishing verifiable public state through the plog.
Suitable custody models include threshold protocols, TEEs, HSMs, OS keyrings, enclaves and other controlled storage systems. The important boundary is that only authorized parties can use private material to decrypt or sign.
BetterSign publishes and verifies public lifecycle state. Private custody remains deployment-specific.
| Threshold keys | Split authority across shares for signing, decryption or recovery |
| TEE or enclave | Restrict key use to hardware-backed execution contexts |
| HSM | Centralized hardware-backed custody for high-value keys |
| OS keyring | Practical local storage for desktop and server workflows |
| MPC | Selected multi-party workflows where key material should not reassemble in one place |