The State of x402

Independent market data on the x402 agent-payments economy, built from our own daily snapshots of the Coinbase Bazaar discovery catalogue. Every figure below is measured, not estimated — including the ones that are unflattering to the ecosystem. Snapshot: 2026-09-15.

30-day GMV
$45,508.06
calls × listed price
Merchants
1,489
Listings
24,685
23,922 with paid traffic
Dead listings
3%
763 earned nothing in 30d
Paid calls (30d)
549,251

Higher prices do not cost you customers

The common instinct is to price at a fraction of a cent to attract agents. The data does not support it. Across every paid band, roughly the same share of listings receives traffic — but the revenue differs by more than an order of magnitude. The $0 band is excluded below because it earns nothing by definition, which makes a revenue comparison against it meaningless.

Price bandListingsGot paid callsEst. 30-day revenue
Micro (≤$0.005)5,816100%$334.34
Cheap ($0.005–$0.05)6,92897%$3,522.47
Mid ($0.05–$0.50)1,97798%$2,258.60
Premium ($0.50–$5)335100%$2,288.41
High ($5+)41100%$37,104.24

What agents actually buy

Ranked by unique paying wallets, not call count. This matters: the loudest endpoints in the catalogue are often a merchant calling itself. A listing with tens of thousands of calls from one or two payers is not demand — it is traffic.

MerchantUnique payersCallsPrice
www.ax1.vc2,00653,821$0.0200
api.onesource.io1,0671,071$0.0030
api.onesource.io9951,001$0.0050
api.onesource.io9721,221$0.0010
api.onesource.io933940$0.0030
api.onesource.io837837$0.0080
api.onesource.io696697$0.0050
api.onesource.io695701$0.0050
api.onesource.io694696$0.0030
api.onesource.io682684$0.0030

Most listings do get paid traffic

763 listings — 3% of the catalogue — took no paid call in thirty days. 546 of them (72%) are priced at $0, against the 39% share that $0 listings hold in the catalogue overall. Among the 15,097 listings that charge anything, 1.4% took no call.

That inverts the usual advice. A $0 price does not lower the barrier to being tried — it reads as an abandoned test entry, and agents skip it. The dead listings we sample share mundane, fixable traits: a blank service name, or a description too thin for an agent to match on. Discovery is not the hard part; being selected is.

Want this for your own listings?

The same dataset powers our paid endpoints: a visibility audit that scores your listings against the catalogue, an authenticity check that separates real demand from self-traffic, and pricing guidance drawn from what actually sells.

Browse the API catalogue →

Method: daily snapshot of the Coinbase Bazaar x402 discovery catalogue, taken once every 24 hours and archived — the discovery API itself keeps no history, so the archive is the only record of how the catalogue changed. GMV is estimated as 30-day call count × listed price and will understate merchants using dynamic pricing. Call and payer counts are self-reported by the discovery API. The API returns one row per accepts entry, so a resource offered on several networks arrives duplicated; listings are deduplicated on (recipient, resource) before any figure on this page is computed, which removes ~3.6% of raw rows. Vibe Springs is itself a merchant in this catalogue and is included in every figure above, including the unflattering ones.