Skip to main content
GET
Index stats
Totals for the whole index, refreshed daily. Costs no credits and needs a valid key, not a balance: this endpoint answers even when your credits are at zero, so an exhausted key can still plan its next purchase. Only the metered endpoints return 402 NO_CREDITS. The counts are materialized once a day rather than counted live, and meta.as_of says when they were taken. Poll it as often as you like; two calls inside a day return the same numbers. Use it to check coverage claims for yourself rather than trusting a number on a marketing page, and to monitor whether the index is growing in the dimension you care about.

Request

Response

object
meta.as_of is when the counts were computed; meta.generated_at is when this response was built. Read as_of when you quote a number.
The counts above are illustrative; call the endpoint for real ones. The match_rates block is the real measured table.

The match-rate table

match_rates is different in kind from the counts around it: a dated measurement of 26 real collections against the index, served as constants so an agent can plan a spend at decision time instead of reading coverage. The chain decides a list’s match rate more than anything else about it, several-fold between Ethereum and Base, so use by_chain and the row for your chain; overall_either_pct describes no collection in particular. either_pct is the planning number: the share of holders with an X or Farcaster account, which is exactly what bills. Pair it with POST /v1/estimate, which reports what your specific list would bill.

Reading the two totals

total_wallets and wallets_checked are different questions and the gap between them is the point. wallets_checked includes wallets we examined and found nothing for. Those are deliberately excluded from total_wallets, so that dividing one by the other gives a real hit rate rather than one inflated by counting empty records as coverage. Do not compute a match rate for your list from these numbers. The index skews toward wallets somebody has already looked up, which is not a random sample of the chain. For your own list, sample it through batch and measure directly. See coverage.

Errors

The standard errors only.