Can an LND seed restore Lightning channels? Aezeed and static channel backups

Learn what an LND aezeed restores, why it is not BIP39, and how a current static channel backup can recover funds without reopening channels.

Can an LND seed restore Lightning channels? Aezeed and static channel backups

No. For a self-operated LND node, the 24-word seed can restore the on-chain wallet, but it does not contain the current Lightning channel state and cannot reopen channels by itself. Off-chain recovery normally needs the same LND cipher seed plus an encrypted Static Channel Backup (SCB). That recovery is an emergency path intended to recover eligible funds on-chain through Data Loss Protection (DLP), not a way to resume the old channels with their latest balances.

This article is limited to self-operated LND. Other Lightning wallets and providers may use different formats and recovery procedures.

What an LND seed actually restores

Aezeed is not BIP39

LND calls its 24-word format an aezeed cipher seed. It uses the familiar default English BIP39 word list, but that shared vocabulary does not make the formats compatible. LND documents aezeed as a versioned, encrypted representation that also records a wallet birthday. Its 24 words encode ciphertext in a format distinct from BIP39. Encryption does not make them safe to share: without a user-selected passphrase, there is no additional secret password protecting them.

Do not paste an LND seed into a BIP39 wallet or assume that a BIP39 recovery screen can derive the same keys. Keep the original LND format and any cipher-seed passphrase. A cipher-seed passphrase is separate from the wallet password: the former protects the seed data, while the latter encrypts the local wallet database. Losing the wallet password is not the same as losing the aezeed, but recovery still requires following LND's procedure.

With the aezeed and its passphrase, LND can recreate the on-chain wallet and rescan for addresses. The rescan depends on the recovery window and the wallet's recorded birthday; an incomplete scan can make funds appear missing even when the seed is correct. That is on-chain recovery. It does not reconstruct the balances, commitment state or pending payments of a Lightning channel.

What a static channel backup contains

An SCB records static information needed to identify a channel and find its peer: for example, the funding transaction, capacity, key-derivation paths, remote node identity and known network addresses. It is encrypted with a key derived from the node's seed, so an SCB cannot be used in isolation. The seed and the backup are complementary pieces of the LND recovery plan.

“Static” does not mean that one file can be ignored forever. LND's channel.backup represents the currently open channels and is updated when channels open or close. Keep the backup copy current, and preserve the latest file after every inventory change. A file from before a channel opened cannot describe that channel; a file that still lists a channel after it closed is also stale for the current inventory.

The SCB is deliberately not a copy of channel.db. The LND safety guide warns that periodically copying a channel database can preserve an unknown or outdated channel state. A stale state can be dangerous because a force-close transaction from an old state may conflict with a later commitment. For the DLP recovery described here, LND recommends its managed SCB, kept current and stored away from the node.

Why an SCB does not reopen a channel

An SCB is not a reliable record of the latest channel balance and commitment transaction. LND's restore process creates “channel shells” with only the information needed to start DLP; those shells cannot be used to route payments as normal channels.

When recovery reaches a peer, LND uses the Data Loss Protection protocol. The peer is asked to force-close the channel and provide the unrevoked commitment information needed to derive and sweep the local on-chain outputs. After the relevant transaction confirms, LND can use the seed and SCB data to recover the keys needed for those outputs. This is a conversion from off-chain channel funds to an on-chain recovery outcome, not a resurrection of the channel.

That path has conditions. The peer must be reachable and respond to the recovery protocol; an unreachable or zombie channel may not be recoverable through SCB. The ordinary documentation also describes recovery for settled funds in the base commitment outputs, not an unconditional recovery of every in-flight HTLC. Fees, timelocks, channel state and the peer's availability affect what happens. An SCB is a safety mechanism, not a promise that every satoshi will be returned immediately.

The dangerous exception is not a normal workflow

The current LND recovery documentation describes a manual force-close option as a last resort when the peer is offline and other routes have been exhausted. Since LND 0.19.0, an SCB or channel-backup response may include an unsigned force-close transaction. Using that material with an outdated channel state is perilous: if the channel was updated after the backup, the other node or a watchtower may broadcast a penalty transaction and seize funds.

Do not turn that warning into a copy-and-paste recovery recipe. It is version-dependent and can have irreversible consequences; the documentation calls it very dangerous and recommends specialist review. Never replay an old channel database or attempt a transaction-executing command merely because a seed and a backup file are available.

A safer LND recovery inventory

  1. Identify the format. Record that the node used LND aezeed, not generic BIP39. Keep the 24 words and, if configured, the cipher-seed passphrase in separate secure locations. Do not photograph or paste them into a web checker.
  2. Preserve the current SCB. Back up LND's channel.backup after channels open or close, and keep a copy away from the node. Never treat an old channel.db snapshot as a safe substitute.
  3. Separate the goal. On-chain recovery uses the aezeed and its passphrase, if configured. The off-chain DLP workflow also needs the encrypted SCB, and its goal is to recover funds through DLP and channel closure, not continue the old channel.
  4. Check the peer and version. Before any recovery, establish which LND version produced the backup, whether the peer can cooperate, and whether a channel is still open. Do not assume a stale or incomplete file covers today's node.
  5. Use the official process privately. Work from LND's current recovery instructions and keep the node's data confidential. If the normal DLP path fails, stop and obtain specialist guidance rather than choosing a manual force close.

Do not test an SCB restore on a live node merely to verify a backup: it can close channels and incur fees. Rehearse in a separate practice environment. If a working node needs moving, follow the official migration procedure; do not run two active copies of the same node.

Where SeedAddress fits

SeedAddress deterministically derives a BIP39 mnemonic in the browser from selected coordinates, an input password, a word list and a word count. Public locations and addresses are not secret random entropy, the method adds no random entropy, and 24 words do not guarantee 256 random bits. SeedAddress does not create an LND aezeed, an LND wallet, a channel database or an SCB, and it cannot recover another node's seed.

For study, use fictional inputs and keep an independent offline backup of the result. Never enter secrets or real node backups into SeedAddress or a web checker. A practice exercise does not verify the backup protecting a live node. For general backup hygiene, see the guide to checking a seed phrase offline; for the difference between a seed and wallet-specific state, read what a seed phrase can and cannot recover.

Sources