What is node?

A computer running blockchain software that keeps a copy of the ledger and relays transactions to other computers.

Nodes independently check that blocks follow the rules, which is what makes a network decentralized rather than trusted. Every wallet, exchange, block explorer, and application ultimately talks to one, either its own or someone else's.

Running a node is not the same as being a validator or a miner, which also propose new blocks. A node verifies and relays. Proposing is an additional role that requires stake or hardware, and most nodes never do it. Conflating the two leads people to think the network is run by a few thousand block producers, when the enforcement of the rules is spread much wider.

How it works

A node downloads blocks from peers, checks every rule for itself, and keeps the resulting state. Rule checking is the important part. If a block producer includes an invalid transaction, or tries to create coins out of nothing, every node rejects the block and it never becomes part of the chain, regardless of how much hash rate or stake stood behind it.

Nodes come in several shapes. A full node validates everything from the current state forward and keeps enough history to serve peers. An archive node additionally retains every historical state, which is what block explorers and analytics need and which costs far more storage. A light client verifies block headers and asks peers for the specific data it needs, which is what makes verification possible on a phone.

Nodes also form the peer-to-peer network that propagates transactions. Your withdrawal is broadcast to a node, held in its mempool, and relayed onward until a block producer sees it.

This is also where the rules of an upgrade are settled. A change only takes effect if node operators run software that enforces it, which is why contentious upgrades are decided by adoption rather than by announcement.

Example

Illustrative comparison of what each kind stores and who runs it.

If you use a wallet without running a node, you are trusting whichever node your wallet connects to, usually one operated by the wallet's developer or a commercial provider.

Why it matters when you buy

For most buyers this is background, but it explains two visible things. Withdrawal delays reflect propagation and confirmation across nodes, and a wallet showing a stale balance is usually a node connectivity problem rather than a lost transaction. It also underpins the reason to verify a deposit on a public explorer rather than trusting a single screen. Per-network measurements are on the chain rankings, and our walkthrough of checking a transaction yourself is at How to use a block explorer.

Related terms

• full node — the standard verifying configuration
• archive node — the history-retaining variant
• light client — verification on constrained devices
• validator — the additional block-proposing role
• rpc provider — the commercial node access most apps use
• mempool — what nodes hold before confirmation

Questions

Do I need to run a node to use crypto?

No. Wallets connect to nodes run by others. Running your own removes the need to trust that connection, at the cost of hardware, bandwidth, and setup.

Do node operators get paid?

No. Running a node earns nothing directly. Operators do it for independent verification, for privacy, or because their business requires reliable access.

How many nodes does a network need?

There is no threshold number. What matters is that nodes are numerous, independently operated, and cheap enough to run that no small group controls whether the rules are enforced.