Hyperliquid node
A Hyperliquid node of your own. Every trade and fill, on your disk.
A Hyperliquid node keeps pace with the chain block by block. Ours runs as a non-validator: it writes trades, fills, order statuses and book changes to files you read directly, and serves a private HyperEVM endpoint. It is sized for Hyperliquid's heavy specs, with a backup node at its side.
Updated . Prices checked .
- Full
- $159/mo
- Ready
- 24 h
On this pageNode type
02Node type
What a Hyperliquid node does, and what it does not.
Hyperliquid has validators and non-validators. Only validators make blocks. A non-validator is the node for reading data.
Full node
$159/mo
- Does
- Follows the chain and writes block data, trades, fills, order statuses and raw book diffs to files
- Serves
- HyperEVM JSON-RPC and a local info server for a subset of requests
- Cannot
- Place orders. Orders still go to Hyperliquid's own API
- Info limits
- No historical time series, no websockets and no l2Book from the local info server
- Archive
- There is no archive node kind. Your history is what the node has written since it started
The one-line test
Do you need every fill and book change as it happens, without rate limits? A node fits. Do you only place orders? The public API already does that, and a node will not.
03How it works
Inside a Hyperliquid node: one program, one upstream peer.
The node streams blocks from a peer, applies them, and writes what it saw to disk.
hl-visor
Hyperliquid's own node software. It checks signed binaries and keeps the node running.
Gossip ports
Ports 4001 and 4002 stay open to the public. Peers favour nodes that accept connections.
Data files
Trades, fills, order statuses and book diffs land on disk. You read them with your server login.
Your endpoint
HyperEVM JSON-RPC and the info server on your nodes.herorpc.com address, behind your key.
A non-validator streams blocks from one upstream peer. If that peer is slow or full, your node falls behind.
The node writes what you switch on: trades, fills, order statuses, raw book diffs, HIP-3 oracle updates.
Hyperliquid documents a lower-latency setting that streams the mempool. It only helps if every peer between you and the validators has it on.
04Know the field
The disk math nobody prints.
The official minimum is a 500 GB SSD. The logs alone outgrow that in less than a week.
Hyperliquid says default settings write around 100 GB of logs a day. The week and month bars multiply that out. Rotating logs is part of running the node.
Non-validator (what we run)
Low-latency setup
Validator (we do not sell)
Checkpoints
A state snapshot is saved every 10,000 blocks. Sizes are from the repository and its open issues, read in October 2026.
CPU, memory and network figures for Hyperliquid nodes, from the official docs: see the node requirements index.
05Hyperliquid, specifically
Six things to know before you run Hyperliquid.
From the node repository and the Hyperliquid docs. Most of them only show up once you run one.
One peer at a time
A non-validator streams blocks from one upstream peer. Operators report "Peer full" errors and nodes stuck at bootstrap when peers are busy.
Two ports must be open
Gossip runs on 4001 and 4002. Peers deprioritise a node that does not accept inbound connections.
Heavy by design
16 vCPU and 128 GB of memory for a non-validator. That is eight times the memory most chains ask for.
Logs grow fast
Around 100 GB a day with default settings. Without rotation, the disk fills in days.
Signed binaries
hl-visor checks the GPG signature of every node binary before it runs it.
The info server is partial
It answers from local state only: no historical time series, no websockets and no l2Book.
06Who rents one
Why teams rent one.
People who need every event, not a polled snapshot.
Market makers and quant desks
Build the book from fills, order statuses and book diffs on your own disk, faster and more detailed than polling.
Data and analytics teams
Record the full stream of trades and fills for research, with no shared rate limit in the way.
HIP-3 market builders
Watch every book and oracle update on the markets you launch, from one node you control.
HyperEVM apps
A private HyperEVM JSON-RPC for your contracts and backends. See every node price side by side.
07Price check
Next to the published prices.
Allnodes lists a dedicated Hyperliquid node, so that bar is like for like. Dwellir only sells clusters now.
Full node
Allnodes: the cheapest dedicated plan on its Hyperliquid pricing page.
Dwellir
Two nodes in a dedicated cluster. A single dedicated node is no longer offered, per its September 2026 post.
You save each month
$161 full
against the Allnodes list price above
08What you get
Set the term.
Pick the node type and how long to prepay. The price updates as you choose.
Your Hyperliquid node
Term
- A dedicated server sized for Hyperliquid
- hl-visor as a non-validator, data files plus one endpoint
- A backup node, included
- A private endpoint on nodes.herorpc.com
- Your server logins
- No setup fee, ready within 24 hours
$159
a month, full node
Billed monthly
Pay in BTC, ETH, USDT, USDC, BNB, SOL, TRX, LTC, DOGE or XMR.
10Questions
Hyperliquid node questions
Real questions from node operators, answered plainly.
What is a Hyperliquid node?
A server running hl-visor that follows the Hyperliquid chain. A non-validator node reads and records data. Only validators make blocks.
Can I trade through my own node?
No. A non-validator cannot place orders. Orders still go to Hyperliquid's API. The node is for reading data faster and in more detail.
What are the Hyperliquid node requirements?
The repository lists 16 vCPU, 128 GB of memory and a 500 GB SSD for a non-validator, on Ubuntu 24.04 only. For the lowest latency, the docs ask for at least 32 logical cores.
Is there a Hyperliquid archive node?
No such node kind exists. Your node keeps what it writes from the day it starts, and state snapshots every 10,000 blocks.
Why is my node stuck or falling behind?
Usually the peer. A non-validator follows one upstream peer, and operators report "Peer full" errors and slow bootstrap when peers are busy.
Does the local info server support websockets?
No. It answers a subset of info requests from local state. Historical time series, websockets and l2Book still need the public API.
Do I get a HyperEVM RPC?
Yes. The node serves HyperEVM JSON-RPC, and your private endpoint passes it through behind your key.
What does it cost?
$159 a month with no setup fee, on a server sized for Hyperliquid. A 12-month prepay takes 25 percent off.
How fast is it ready?
Inside a day of paying, often sooner. Once it has caught up, the endpoint URL and SSH details sit in your dashboard.
Every fill, on a node only you read.
$159 a month, a backup node included, live inside a day.