Why x402 numbers disagree
Trackers report different x402 numbers mainly because they count different things. In the 7 days to 2026-10-15, Base x402 payments totalled $180,899. Our first four screens leave $176,007 (97%); the two concentration screens (single-buyer merchant, one buyer ≥90%) then leave $23,975 (13%).
Different trackers measure different windows, chains and things. None of the figures below is “wrong”; each answers a different question. This page puts them side by side, with sources and dates, and shows step by step how our own number is built.
What each tracker counts
| Tracker | What it counts | What it leaves out | Window | Headline figures (as read) |
|---|---|---|---|---|
| Agent Pay Tracker this site, 2026-10-15 | x402 settlements sent by registered facilitators, decoded wallet by wallet. | Payments tripping any of six screens, including two merchant-concentration screens. | 7 days Base and Solana | Base: reported $180,899, passed $23,975 (13%) Solana (sampled): reported $87,662, passed $31,997 (37%) |
| x402scan read 2026-10-15 | Every x402 settlement it indexes across chains: transactions, volume, buyers and sellers. | Nothing; it reports raw totals. | Past 30 days (as shown on its homepage) All chains it indexes (Base, Solana) | Transactions, 30 days: 2,888,990 Volume, 30 days: $1,095,102 Buyers: 36,615 Sellers: 38,887 |
| x402watch read 2026-10-17 | x402 services indexed from the Coinbase Bazaar feed; each buyer gets one of 8 labels (e.g. organic_user, ai_agent, developer, self_test, suspected_wash). | Real volume excludes buyers labelled self_test, suspected_wash and developer; owner_test traffic is left out of the denominator. | Rolling 30 days (daily CC0 snapshot) Indexed services (EVM and Solana) | Real volume share, 30 days: 90% Real volume, 30 days: $548,818 Total volume, 30 days: $610,603 Suspected wash volume, 30 days: $913 Developer volume, 30 days: $9,648 Transactions, 30 days: 730,543 |
| Artemis read 2026-10-17 | x402 and MPP (Machine Payments Protocol) payments across chains, shown as cumulative and 'Adjusted' series by chain, facilitator and category. | Its adjusted series and 'Real vs. Gamed Transactions' chart separate genuine demand from wash trading; the exact rules are described by Artemis, not reproduced here. | Cumulative and time series Multiple (x402 and MPP) | Figures are rendered in interactive charts; the page text does not state a single headline number, so we don't quote one. |
| Dune (x402 Facilitators dashboard) read 2026-10-17 | Settlements sent by known x402 facilitators, per facilitator and chain, all time. | Nothing; it reports raw facilitator activity. | All time (from May 2025) Base, Polygon and others | PayAI on Base, all-time settlements: 18,312,035 PayAI on Base, all-time volume: $4,405,418 |
| x402 List read 2026-10-17 | On-chain settlement volume routed through the facilitators it tracks, plus uptime of about 1,000 listed services. | No real-vs-wash filter on the settlement figure. | Rolling 30 days Multiple | Settlement volume, 30 days: ~$1,330,000 Settlements, 30 days: 6,642,963 |
Competitor figures last checked: 2026-10-17. Hand-checked entries (Artemis, Dune, x402 List) are re-checked monthly; x402scan (2026-10-15) and x402watch (2026-10-17) update automatically with each weekly refresh.
Competitor figures are quoted as published on the date shown, with a link to the source. x402scan figures come from its public homepage; x402watch figures from its CC0 daily data snapshot on GitHub. If you run one of these trackers and something here is out of date, tell us on X and we'll fix it.
From reported to real: Base, 7 days to 2026-10-15
Screens are applied in this order; each payment is counted once, at the first screen it trips.
| After screen | Payments removed | Volume removed | Volume left | Share left |
|---|---|---|---|---|
| Reported (all x402 payments) | — | — | $180,899 | 100% |
| Self-payment | 24 | $0.05 | $180,899 | 100% |
| Circular flow | 931 | $305 | $180,594 | 100% |
| Buyer funded by the merchant | 49,423 | $4,040 | $176,554 | 98% |
| Tiny repeated payments | 172,486 | $547 | $176,007 | 97% |
| Single-buyer merchant | 4,609 | $144,574 | $31,433 | 17% |
| Buyer concentration | 14,643 | $7,458 | $23,975 | 13% |
From reported to real: Solana, 7 days to 2026-10-16 (sampled)
| After screen | Payments removed | Volume removed | Volume left | Share left |
|---|---|---|---|---|
| Reported (all x402 payments) | — | — | $87,662 | 100% |
| Self-payment | 49 | $2.78 | $87,659 | 100% |
| Circular flow | 4,210 | $398 | $87,261 | 100% |
| Buyer funded by the merchant | 1,384 | $344 | $86,917 | 99% |
| Tiny repeated payments | 44,609 | $78.06 | $86,839 | 99% |
| Single-buyer merchant | 2,099 | $57.50 | $86,782 | 99% |
| Buyer concentration | 5,771 | $54,785 | $31,997 | 37% |
Sampled: figment — 5,000 of 173,199 transactions decoded at random. Sampled payments are counted once (not scaled up), so Solana totals are a lower bound.
Why the gaps
- Window and coverage. x402scan and x402 List report 30 days across the chains they index; Dune reports all time per facilitator; we report 7 days on Base and Solana. A 30-day multi-chain total will be several times our 7-day figure before any screening.
- Labels vs payment screens. x402watch labels each buyer and excludes self-test, suspected-wash and developer traffic; its CC0 snapshot of 2026-10-17 gives a 30-day real share of 90%. Our first four screens are closest to that idea, and on Base they leave 97% of volume (Solana: 99%). The two figures sit in a similar range.
- Merchant concentration. Most of the difference comes from two screens: single-buyer merchant and buyer concentration (one wallet paying 90% or more of a merchant's volume). On Base they remove $152,032 of $180,899. We exclude this volume because one wallet paying one merchant doesn't show a market of agents buying; other trackers can reasonably count it.
- Adjusted series. Artemis publishes adjusted (“real vs. gamed”) x402 and MPP series in interactive charts. We don't quote a figure because the page text doesn't state one.
- Payments, not proof of agents. Passing every screen means no red flag in our checks; it does not prove a payment was made by an AI agent.
Full rules: Methodology. Data: /api/v1/compare.json.