> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hyperdash.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Hyperliquid Peering

> A private, whitelisted peer connection between your Hyperliquid node and Hyperdash infrastructure.

[Contact](https://t.me/hyperhans)

Teams running their own Hyperliquid nodes depend on peers for market data. Public gossip peers are capacity-constrained, disconnect without notice, and degrade under load, particularly around network upgrades.

> [!NOTE]
>
> **Hyperliquid Peering is a private, whitelisted connection between your node and Hyperdash infrastructure, which peers directly with a Hyperliquid Foundation node.** No middle layers between the Foundation peer and your node.

## Who it is for

Teams that already run, or are ready to run, their own non-validating Hyperliquid node and want a stable path to the network:

* Market makers and latency-sensitive trading firms
* Quant and research funds
* Data and analytics providers
* Infrastructure teams building on Hyperliquid

## Peering or a hosted feed

Peering removes the distribution layer entirely. Any hosted feed has to receive, process, and rebroadcast data before it reaches you. With direct peering, your node receives blocks over Hyperliquid's native gossip protocol and you decode locally.

Peering suits teams already running their own node. If you would rather not operate one, talk to us about API access instead.

## What you get

* **Blocks and state updates** over Hyperliquid's native gossip protocol
* **Mempool transactions**: our infrastructure runs `split_client_blocks` end to end, so `mempool_txs` is populated on your node
* **Three whitelisted IPs** included, reserved for your nodes rather than shared capacity
* **Further IPs** on request, so you can move from staging to production without re-onboarding

We are never in your execution path. Order submission goes directly to Hyperliquid's API.

## Regions

| Region | Location | Status |
| - | - | - |
| Japan | Tokyo | Active |
| Europe | Germany | Active |

Place your node in the same region as the node you peer with. Fewer hops mean lower tail latency to fills, blocks, and state updates. Actual improvement depends on your own infrastructure, region, and processing pipeline, and gossip-layer latency is set by the network.

## Before you start

You will need a non-validating Hyperliquid node that you operate, on a machine with a static public IP. Hardware, operating system, and port requirements are set by Hyperliquid, not by us: see the [Hyperliquid node repository](https://github.com/hyperliquid-dex/node) for current specifications.

Size disk and bandwidth for mempool capture before you enable it. `mempool_txs` is written continuously and is significantly larger than block data alone.

**Recommended starting point on AWS:** the x8aedz instance family, with 5th-gen AMD EPYC (Turin) processors at 5.0 GHz, the highest clock on AWS, and local NVMe storage. It runs in Tokyo, the same region as our Japan node. Start there and tune for your own workload.

## Setup

Setup is completed the same day.

**1. Whitelist**

Send us the static IP of your node. We whitelist it on our node and return our node IP for your region.

**2. Configure**

Set our node IP as the root entry in your override gossip config, at `~/override_gossip_config.json`:

```json theme={null}
{
  "root_node_ips": [{ "Ip": "<node-ip-we-provide>" }],
  "try_new_peers": false,
  "chain": "Mainnet"
}
```

Setting `try_new_peers` to `false` keeps your node on the reserved peer instead of drifting back onto public ones.

**3. Activate**

Restart your node and confirm blocks are applying.

## Resources

> [card] [Hyperliquid node setup](https://github.com/hyperliquid-dex/node)
>
> Official node software, hardware requirements, and setup steps.

> [card] [Mempool streaming](https://github.com/hyperliquid-dex/node#streaming-mempool-transactions)
>
> How to stream `mempool_txs` from your node.

> [card] [Priority fees](https://hyperliquid.gitbook.io/hyperliquid-docs/for-developers/api/priority-fees)
>
> How Hyperliquid's priority fees work.

> [card] [Hyperliquid API on Telegram](https://t.me/hyperliquid_api)
>
> Hyperliquid's channel for API and node developers.

## Uptime and support

* **99.99% uptime**, with 24/7 monitoring and alerting
* Direct channel to our infrastructure team

## Trial period

A complimentary trial lets you benchmark against your existing setup from your own machines.

## Pricing

**\$500/month** or **\$5,000/year**, with three whitelisted IPs included. Message us on [Telegram](https://t.me/hyperhans) to start a free 7-day trial.

Staking HYPE to our validator qualifies for discounts.

## FAQ

> [accordion] How does this differ from consuming a websocket feed or API?
>
> Peering removes the distribution layer entirely. Any hosted feed must receive, process, and rebroadcast data before it reaches you. With direct peering, your node receives blocks over Hyperliquid's native gossip protocol and you decode locally.

> [accordion] What latency improvement should we expect?
>
> Results depend on each firm's infrastructure, region, and processing pipeline, and gossip-layer latency is set by the network. The trial period exists so you can measure it from your own machines rather than take a number on trust.

> [accordion] Where does our node get its data?
>
> Our node peers directly with a Hyperliquid Foundation node. It originally peered with our validator infrastructure, which was the lowest-latency path at the time. Since Hyperliquid changed its gossip architecture, direct peering with the Foundation node is the fastest path. That Foundation peer is dedicated to our node.

> [accordion] How does peering compare to public peers?
>
> Public gossip peers are capacity-constrained and disconnect without notice, particularly around network upgrades. Your peer here is reserved for your node.

> [accordion] Is mempool data included?
>
> Yes. Our infrastructure runs `split_client_blocks` end to end, so `mempool_txs` is populated on your node.

> [accordion] Are you in our execution path, and is our activity private?
>
> No. Order submission goes directly to Hyperliquid's API, and we are never in the execution path. We do not record or retain data that could expose your activity.

> [accordion] What is the uptime commitment?
>
> 99.99% uptime, with 24/7 monitoring and alerting.

> [accordion] Can we change or add whitelisted IPs?
>
> Yes. Three are included, and they can be swapped as you move from staging to production. Further IPs can be added on request.

> [accordion] How long is the trial?
>
> 7 days.

> [accordion] Which AWS zone is the Tokyo node in?
>
> AZ ID `apne1-az1` in ap-northeast-1.

> [accordion] Does peering speed up order submission?
>
> No. Orders go through Hyperliquid's API; peering speeds up data only.

> [accordion] Why can't our node reconnect after several restarts?
>
> Hyperliquid allows each IP one sync request every 10 minutes. Wait 10 minutes and your node will reconnect.

> [accordion] How does billing work?
>
> We offer monthly (\$500) or annual (\$5,000) plans.

> [accordion] Does any client get priority over others?
>
> No. Every client gets the same connection and priority. We do not enable gossip priority auctions on our nodes.

## Access

To get access, email [sales@hyperdash.com](mailto:sales@hyperdash.com) or reach us on [Telegram](https://t.me/hyperhans).

## About Hyperdash

[Hyperdash](https://hyperdash.com) is a global trading platform and data analytics company. Hyperliquid Peering runs on our production infrastructure.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.