List your API. Get paid when it delivers.

Joining is an application rather than a listing. What we need from you is the set of things about your data that cannot be observed from outside: who you are, where the data comes from, what your endpoint does when your own upstream fails, and on what terms you sell it.

What we never ask you

There is no field for latency, uptime, accuracy or error rate. Not because they do not matter, but because they matter too much to take on trust. Metera measures those itself, continuously, from real traffic, and publishes what it measures.

A form that let a provider grade themselves would turn the network's central claim into a self report. The one thing we do ask about performance is permission: consent to be measured and published is a condition of being listed, and it is the only item in the application that cannot be negotiated.

What we do ask

Identity and jurisdiction, so we know who stands behind the data and which rules apply to it. Then which catalogue types you serve and what you cover for each, read live from the running catalogue rather than from a fixed list.

Provenance, which is the section that matters most to the engine. Whether you originate, aggregate or relay, and which upstreams you read. Two providers reading a single upstream are one observation counted twice, and consensus between them would be a number reported twice rather than confirmed. Naming your upstreams does not count against you. Withholding them makes your data worth less to us, because we then have to assume the worst about independence.

Behaviour under failure. Whether you push or we pull, the heartbeat you guarantee if you push, and what your endpoint returns when your own source goes down. Returning an error is the clean answer. Repeating the last value is the one that needs explaining, because a stale number that still looks fresh is the exact failure this network exists to catch.

Then the commercial terms: pricing, rate limits, redistribution, attribution, data rights, and any regulatory restriction on who you can serve.

What happens after you apply

Your endpoint is probed automatically. We check that it answers and how quickly, whether the live response carries the keys your sample declared, whether the TLS certificate is valid, and whether the domain matches the organisation you named. Then the cheapest and most useful check of all: whether your first reading agrees with what the network already measures for the same pair.

A person then reads the application beside what the probe found, side by side, with any disagreement marked. A declared thirty second heartbeat against a response that took four minutes is a question rather than an automatic rejection, but it is a question you will be asked.

Approval moves you to sandbox rather than straight to production. We measure you there against the sources already running, without routing live traffic to you, until there is enough evidence to rank you honestly. Promotion into production is gradual after that.

How you earn

On calls that deliver and pass verification. Routing is blind to price and to stake, so nobody can pay to be picked and nobody can pay to rank higher. What you get by joining is traffic at the moments you are genuinely the best source available, and a public record of how you actually perform.

A refusal costs the caller nothing, and a source that fails verification is swapped inside the same request. Both of those are the same bet from your side: the better you actually are, the more of the traffic you get, with no sales conversation in between.

Applying

Sign in first. The application is five steps and saves as you go, so you can leave it and come back to it, and it lives with your account rather than behind an anonymous link. That account is also where your sandbox keys arrive if you are approved.