Terms of use

Every figure describes what happened between one of our probes and one endpoint, over a stated window, at a stated sample count. That is a narrower claim than it may look like, and the difference matters.

• Latency is a property of a network path, not of a provider. A figure measured from one region describes that region. Your users are somewhere else.
• The past is not a forecast. A provider that was fast for thirty days may be slow tomorrow. Nothing here predicts.
• A free endpoint is not the paid product. Most endpoints measured here are public tiers. A provider’s commercial service may behave completely differently, and its SLA usually does not cover what we measured.
• Absence is stated, not filled in. Where we have too little data we say so. We do not estimate, and a blank is a blank rather than a zero.

The full method, including every threshold, window and scoring weight, is published at /methodology. If you intend to make a decision on these figures, read it first.

Nothing here is financial, investment, legal or operational advice, and nothing here is a recommendation to use or avoid any provider, chain, validator or staking arrangement.

The staking pages report what was paid in the past, recovered from what each chain publishes. They are not a projection of what you would earn, they exclude MEV entirely, and they say so on the page. Staking carries risks — slashing, lockup, validator failure, and the price of the asset — that no yield figure addresses.

Rankings are never influenced by payment. There is no paid placement and no advertising on any page that carries a measurement.

You may read, quote, cite and build on what is published here, including through the public API. We would rather the figures were used than hoarded, which is why every table exports and every endpoint is documented and unauthenticated.

Two conditions, both about not misrepresenting what a figure is:

• Carry the context. A figure reproduced without its window, its sample count or its vantage point is a different and weaker claim than the one we made. Cite the page it came from.
• Do not present it as ours. Do not attribute conclusions to RPCNode.IO that RPCNode.IO did not publish, and do not imply that we endorse a product, provider or service.

The API is offered at conservative rates and without a guarantee. Please cache rather than poll; if your use is heavy enough to matter, write to us and we will work something out rather than rate-limit you by surprise.

• Do not attempt to gain access to any part of this system you were not granted — the admin console, another account, or any host behind it.
• Do not use the API in a way that degrades it for others: scraping at a rate that displaces normal readers, or automated retries that ignore an error response.
• Do not use anything obtained here to attack, overload or disrupt any measured provider’s infrastructure.
• Do not misrepresent a figure in a way calculated to damage a provider — for example by republishing an outage without the window it occurred in.

Found a vulnerability in this platform? Write to the security address below. We would much rather hear from you than read about it, and we will not pursue anyone who reports in good faith and gives us a reasonable chance to fix it before disclosing.

The site and the API are provided as they are, with no warranty and no uptime commitment. We run a single deployment; an outage costs collection continuity, and where collection stopped the site reports a gap rather than papering over it.

Measurement methods change as they are improved. Where a change alters what a figure means, the change is published and the version that produced each observation is stored with it, so a figure computed under one method is never silently compared with another.

These terms may change. The date at the top of this page moves when they do, and it moves only for changes to the terms rather than for corrections to spelling.

To the extent the law allows, we are not liable for any loss arising from use of this site or its data — including any decision to use, avoid, delegate to or rely on any provider, validator or chain.

We take care to publish only what we measured, to state what each figure covers, and to correct what turns out to be wrong. That is the commitment we make, and it is not the same as a guarantee that every figure is right.

Nothing in these terms limits liability that cannot lawfully be limited.

hello@rpcnode.io — general enquiries

providers@rpcnode.io — if we are measuring your infrastructure

security@rpcnode.io — vulnerability reports

privacy@rpcnode.io — data protection requests