Hyperliquid Peering

Dedicated, SLA-backed gossip peers for the Hyperliquid node you run yourself.

You keep operating your own node. Dwellir reserves a peer slot for it, so your node keeps a stable upstream instead of competing for public root peers.

What you get

Why peering matters

Your Hyperliquid node keeps one peer connection that actually feeds it. You live and die by that peer's throughput and uptime. If it is slow, your node is slow. If it drops, your node stops.

The community root peers listed by the gossipRootIps API are shared and oversubscribed, and the one you land on is a dice roll. Their addresses change and their slots fill, so the peer that worked yesterday may not take you today.

When your node loses its peer it stops advancing, falls behind the tip, and restarts sync. Downstream that means stale order books, missing fills, and gaps in whatever you write to disk.

A reserved slot removes the dice roll. Your node's IP sits in reserved_peer_ips on both Dwellir peers, so a slot is held for you whether or not the network is busy, and you know which machine you are pulling from.

Requirements

Before you subscribe, confirm the following about your node:

1. Ports 4001 and 4002 are open to the public internet. Hyperliquid uses them for gossip. A node that does not accept inbound connections on these ports gets deprioritized by every peer in the network, including ours.
2. You have a static public egress IP. Dwellir allowlists that exact address. Dynamic IPs and large NAT ranges do not work.
3. You run hl-visor in non-validator mode on Mainnet.

Setup

1. Request access

Email support@dwellir.com or send a Telegram DM to @dwellirgrowth with:

• The public IP of each node you want to peer
• The region your node runs in

2. Receive your peer IPs

Dwellir adds each node IP to reserved_peer_ips on both peers and returns two peer addresses.

3. Point your node at them

Write the peer IPs into ~/override_gossip_config.json:

{
  "root_node_ips": [{ "Ip": "203.0.113.10" }, { "Ip": "203.0.113.11" }],
  "try_new_peers": false,
  "chain": "Mainnet",
  "reserved_peer_ips": []
}

Replace the example addresses with the IPs Dwellir sends you.

Keep try_new_peers set to false. With true, your node discovers and connects to arbitrary peers off the public network, which puts you back on the dice roll you are paying to avoid. With false, it talks to the Dwellir peers and nothing else.

If you need the mempool, also set split_client_blocks to true so the node surfaces uncommitted transactions.

4. Restart and confirm

Restart hl-visor, then confirm the node keeps applying blocks. The node writes a new file per block batch under its data directory, so the newest file should be seconds old:

find ~/hl/data/replica_cmds -type f -newermt '-1 minute' | tail -5

If the newest file stops advancing, or the visor logs show repeated peer disconnects, send the logs to Dwellir support and the reserved slot gets checked from the peer side.

Service level

Uptime: 99.9% per calendar month, per peer. Dwellir measures gossip reachability on ports 4001 and 4002 from outside its own network. Scheduled maintenance is announced at least 48 hours ahead and is excluded from the measurement.

No preferential treatment. Every peering subscriber gets the same peers, the same configuration, and the same priority. Dwellir's own infrastructure runs on these same peers, on the same terms. Dwellir does not run a faster path for market makers, for itself, or for any individual customer. If you have evidence otherwise, email support@dwellir.com. Dwellir will investigate and publish the result.

Pricing

$500/month flat, per allowlisted node IP.

• 2 Dwellir peers to connect to, single region
• Full mempool gossip
• 99.9% uptime SLA
• No usage, bandwidth, or egress charges
• Monthly billing, cancel any time

Prefer Telegram? DM @dwellirgrowth.

What peering does not include

Peering gives your node a reliable upstream. It does not give you data services.

Those are powered by the client the Dwellir team built, which streams the raw files a Hyperliquid node writes to disk over gRPC, reconstructs order books, and indexes history. If you would rather consume that than run and maintain a node, start here:

Frequently asked questions

Can I peer more than one node?

Yes. Pricing is $500/month per allowlisted node IP, so two nodes cost $1,000/month. Each node needs its own static egress IP.

What happens if I change my node's IP?

Send the new address to Dwellir support, either by email or on Telegram at @dwellirgrowth, and keep the old one running until the change is confirmed. With try_new_peers set to false, an unallowlisted node has nothing to fall back to and will stop syncing.

Do you throttle the mempool stream?

No. You receive full gossip, the same stream Dwellir's own nodes consume.

Is peering required to use Dwellir's Hyperliquid APIs?

No. The gRPC streaming API, Order Book WebSocket, and Info endpoint all run on Dwellir infrastructure, and that infrastructure sits on the same peering backbone offered here.

Buy peering only if you run your own nodes and want more reliable peers. If you just want access to Hyperliquid data, use the Dwellir API platform or a Dedicated Cluster instead.

Who do I contact?

Email support@dwellir.com, send a Telegram DM to @dwellirgrowth, or go through your existing Dwellir contact. Subscribers also get a shared Telegram or Slack Connect channel with the team.