RPC

A full node serves the public rpc.Api service from its local replicated state. Use it as a private or dedicated RPC endpoint for applications, indexers, explorers, and other infrastructure that should not depend directly on official public RPC capacity.

Local Queries

Read queries are served from the full node's local state. This includes account, block, note, and sync methods.

The network-note debugging endpoint, GetNetworkNoteStatus, is the exception: it depends on NTX builder state rather than replicated chain state, so full nodes forward it to the configured upstream RPC source.

Transaction Submission

Full nodes do not sequence transactions. SubmitProvenTx and SubmitProvenTxBatch are forwarded to the configured upstream RPC source. If the upstream source is another full node, the request is forwarded again until it reaches the sequencer.

This lets clients submit through a nearby full node while preserving the sequencer as the only block-producing node.

Scaling Throughput

Full nodes sync from their upstream source using:

• BlockSubscription for committed signed blocks.
• ProofSubscription for block proofs.

You can daisy-chain full nodes to scale read throughput, and fan out.

                             ┌── Full node
Official RPC ── Full node ───┼── Full node
                             └── Full node

Each downstream full node subscribes to blocks and proofs from its upstream source, stores local state, and serves its own read RPC traffic. This can reduce load on the sequencer or official public RPC endpoint.

See the RPC subscriptions guide for stream semantics and reconnect behavior.