x402 on-chain receipts — the discovery catalog, checked against the chain

The CDP x402 discovery catalog publishes calls_30d per resource. Gross = advertised price × calls is not settled money. This page counts actual USDC Transfer events to each payTo and reports the ratio.

471ax1.vc transfers sampled (Base)
$0.02every one of them, exactly
96distinct payers, ax1.vc
3median sender nonce
$622,851.94bitrefill payTo, Arbitrum, full 30d

Base — 24 windows × 50 blocks, spread over 30 days

2400 s sampled = 0.0926% of the month. Detection floor ≈ 36.0 transfers/day: a zero here only rules out daily volume above that floor. Head block 52655200.

resourcepayTosample txsenderssample USDCon-chain est. /30dcatalog /30dratio
cheaptokens.ai0x213c8d74…32152250.980055,058.40492.00111.91×
www.ax1.vc0x7284D41b…299C464979.280010,022.406,991.381.43×
stableenrich.dev0x325bdF6F…d4302951.90722,059.78281.407.32×
keyring-agent.blockchhub.lin0x17cd53C0…790518160.9000972.001,409.050.69×
laso.finance0x3291e96b…2b37110.00000.02260.000.00×
api.bitrefill.com0x480CD46E…846A000.00000.0051,000.000.00×
agi.apify.com0x4aAbE17C…DF26000.00000.00——
CONTROL_0x515d935b0x515d935b…9e11000.00000.000.00—

ratio = on-chain estimate ÷ catalog figure. green = within 2×, orange = chain above catalog, grey = chain below.

ax1.vc — who is actually paying

What this does and does not prove. It proves USDC really moves to this payTo at the advertised unit price, from 96 distinct addresses, at the catalogue's order of magnitude. It does not prove those addresses are independent economic actors — a self-operated fleet of fresh wallets looks identical under this method, and this payTo was already falsified as self-farmed on 2026-10-03 by a stronger test (6 consecutive 2 h windows: 64 payers, 35 appear in one window only, none in more than 3 of 6; 12 wallets share a byte-identical per-window count vector; 5 sampled payers send 100% of their outbound USDC to this address and nothing else; one cohort stops in the same block). Today's independent sample is consistent with that verdict: low nonces, ~0 native balance, bursty windows, zero cross-purchasing.

So the headline is not "x402 is a $10k/mo market". It is: the catalogue's advertised volume for this resource tracks real on-chain transfer volume within 1.4×, and that on-chain volume is a self-operated farm. A catalogue row can be numerically honest and economically empty at the same time.

bitrefill — the catalog's #1 row, off-Base

Catalog: 51 calls × $1,000 = $51,000/30d (81.5% of the whole catalog's gross). On Base, in sample: 0 transfers. On Arbitrum, one query over the full 30 days:

Caveat: a gift-card issuer's deposit address is very likely a general treasury, not an x402-only payTo. Do not read $622k as x402 revenue. The defensible claim is narrower: the catalog's advertised-price × calls metric is not a proxy for settled volume in either direction — it understates here by 12× and was never checked at all.

Method — what each chain actually allows

Reproduce