Real leaderboard

Every x402 leaderboard on the internet ranks by the same quantity: total settled volume. It is the honest thing to publish if you only index settlement, because a payment that clears on-chain looks identical whether two independent parties made it or one wallet moved money to itself in a loop. Rank by that number and the top of the board belongs to whoever generated the most transactions — which is not the same question as who sold the most.

So we publish both rankings. The left-hand rank on every row below is the raw one, on the same all-time verified basis the rest of the market quotes. The right-hand rank is what is left after payments we can convict as non-independent are subtracted, edge by edge, under a named rule. Where a name barely moves, the raw board was telling the truth about it. Where a name falls dozens of places, it was not.

The subtraction is deliberately conservative. Only edge-level verdicts are removed — an address paying itself, a pair cycling funds, a funder fanning out to payees that sell to nobody else. Week-level concentration, which would flag most of the market, is measured but published separately as a bound, never subtracted here.

Net-of-wash leaderboard

Board not published yet — the mart lands on the next rebuild.

How to read this

A large fall is not an accusation.

It says a large share of the volume published under that name is not two independent parties transacting. Testing, liquidity bootstrapping and internal accounting all produce exactly that shape and all are legitimate. The narrow claim is that it should not be counted as market demand.

Seller haircuts are measured; facilitator haircuts are attributed.

A seller's own edges carry the verdict directly. A facilitator only relays, so its figure is each flagged seller-day apportioned pro-rata to the signers that processed it — an estimate, and a lower bound, since flagged volume no known facilitator processed is attributed to nobody. Each row states which basis it is on.

“Not screened” is not “clean”.

A row with no wash run behind it shows no bar and no adjusted rank. A row that was screened and came back with nothing shows a measured 0%. Those are different findings and never share a pixel here.

Some payees have no name.

A payee with verified settlement that our catalog has not matched to a service shows as Unidentified seller. It keeps its rank — the volume is real and excluding it would flatter the board — but its internal key is never printed.

Check the workings

• →Integrity — the same subtraction at ecosystem scale, per chain and month, plus the concentration bound that is measured but deliberately not removed.
• →All services — the full directory on the raw basis, with per-service detail, trust tiers and payment rails.
• →Facilitators — the relayer side on its raw basis, with signer registries and evidence status. The raw totals here are that page's, rolled org-wide.
• →Coverage — per rail, what was indexed and how it was verified, so the raw basis itself can be audited before anything is subtracted from it.