Skip to main content

Running a node

Anyone can run a node. No permission, no stake required.

The operational guides live with the code so they stay correct as the software changes:

  • Running a node — binary, genesis and its checksum, seeds, gas price, pruning, hardware.
  • Upgrades — how a coordinated upgrade runs, and what goes wrong.
  • Releases — binaries and checksums.

What is different about this chain​

Every registration verifies a zero-knowledge proof on-chain. That is unusual, and it is CPU-heavy.

Two consequences:

  • Budget more processor than a chain of this size would normally need.
  • Use state sync rather than replaying from genesis, because replaying re-verifies every proof ever submitted. Sync cost grows with adoption, not just with time.

Becoming a validator​

There is no allowlist. Acquire ERTH, self-delegate, and submit MsgCreateValidator.

There was no genesis allocation for validators. ERTH is earned — by registering a passport, by bidding in the liquidity auction, or on the market.

Keep your consensus key off the node itself. The join guide covers the remote signer setup, which fails closed: with a signer configured and none answering, your node signs nothing rather than signing with a key it should not hold.