Endpoint scores

A provider score tells you about a company. It cannot tell you which of that company's endpoints is the one worth calling, and some publish thousands. This board scores each endpoint on its own evidence: whether it answers when we knock, how quickly, whether it says what it does, and whether its price and payment address stay put. The ranking blends that endpoint score (80%) with the provider's own trust score (20%) — both are shown separately per row, and an unscored provider never counts against its endpoints. It shows the strongest of them — up to 120 rows, at most three per provider domain and one per parameterized route — so one large provider cannot fill the page.

Payment history is deliberately absent from this number. Almost every endpoint shares a payment wallet with its neighbours, so on-chain activity belongs to the provider rather than to any one endpoint. Rows that do have their own wallet are marked, and the provider's own trust score stays where it was, on Services.

A dash is not a zero. It means we never measured that pillar, and the page says so rather than filling the gap with a number.

Related

• →Catalog — the same endpoints browsed by what they do, with each price scored against its category peers.
• →Services — one row per provider, ranked by verified settlement volume and scored on provider-level evidence.
• →Endpoint stats — how much of the endpoint fleet we probe, how often, and what the probes find.