Every x402 seller we can crawl, ranked by real Base USDC settled on chain - not self-reported traffic, not a press release. Read from event logs, refreshed hourly, and free to query.
Agent402 is excluded from this ranking by default. An index that puts itself first is not evidence of anything, so our own figure is published separately below.
Partial scan - 7d scan, partial: 1 of 204 ranges unavailable; totals are a floor, not the full window.
| sellers ranked | 1,060 |
|---|---|
| wallets queried | 1,060 |
| bazaar listings | 15,134 |
| blocks scanned | 302,400 |
| window | 7d |
| per-call ceiling | $0.75 |
Volume alone can be manufactured by paying yourself, so every row also carries distinct paying wallets and an organic ratio of buyers to calls. A thousand calls from two wallets reads very differently from a thousand from four hundred.
We run the index, so we keep ourselves out of the ranking above. Here is the same figure for us, from the same public endpoint. It counts paid calls in stablecoin only, excluding free proof-of-work calls and our own monitoring probes.
| calls paid in stablecoin | 28,878 |
|---|---|
| free over proof-of-work | 8,481 |
| rails with settled traffic | 12 of 12 |
| settled over the MPP wire | 87 |
Most sellers here settle on Base alone. Twelve rails is the difference, and it is checkable on chain.
What this cannot see: settlements on chains other than Base, payments to addresses a seller never advertised, and which specific tool was bought. Those are limits of on-chain data, not of the crawler.
Sellers are discovered from the Coinbase CDP Bazaar plus Agent402's own crawl, their payTo addresses are read from their x402 manifests, and settlement transfers to those addresses are aggregated from Base event logs via eth_getLogs. A per-call USD ceiling excludes transfers too large to be a single tool call, so ordinary treasury movements do not inflate a seller. The snapshot refreshes hourly and the raw JSON is free at /api/leaderboard.
The table above excludes our own wallet, the same filter /api/leaderboard applies with include=external. An index that ranks itself first is not evidence of anything, so the neutral view excludes the operator by default and our own figure is published separately, from the same public endpoint.
Volume alone, yes - paying yourself inflates totalUsd for free. That is exactly why each row also carries unique buyers and an organic ratio of buyers to calls. A seller with thousands of calls from two wallets reads very differently from one with hundreds of calls from hundreds of wallets, and both are visible here rather than averaged away.
Serve x402 challenges, register your origin, and the crawler picks you up on the next pass. Free, no signup, nothing deducted from your price.