What goes in, how it is measured, what is excluded
x402.be aggregates the public Bazaar APIs published by x402 facilitators. Today that means the CDP Bazaar (api.cdp.coinbase.com) and the PayAI Bazaar (facilitator.payai.network). Both are read-only, key-free, and rate-limit free at the time of writing. Other facilitators (SolPay, AutoIncentive, Dexter) will be added when they publish public discovery endpoints.
Each Bazaar item carries a 30-day cumulative call count and a 30-day unique-payer count, self-reported by the facilitator. Volume is computed as price × calls — where price is the first amount in the resource's accepts[] array, converted from base units to USD assuming 6-decimal USDC.
The Bazaar APIs do not yet expose daily breakdowns or individual payment events. Until they do, all figures on this dashboard are 30-day cumulative. We label this everywhere ("30d") rather than synthesizing 24h windows from monthly totals.
Data is fetched on demand and cached for 60 seconds via Next.js ISR. The "live · sources · Ns ago" pill on each page shows the actual age of the snapshot. If both Bazaar sources fail, the dashboard renders a deterministic mock fallback and the pill switches to "fallback".
Multichain by design. Solana, Base, Optimism, Arbitrum, Polygon, Ethereum, Avalanche, World, Monad, Stellar, Sei are all first-class; any chain a known facilitator advertises in accepts[].network is automatically picked up. Testnets are folded into their mainnet bucket. No chain is preferentially weighted in rankings.
Volume is attributed to the FIRST chain a server lists in accepts[]. A multichain server that accepts Base and Solana shows up under Base if Base is listed first. This means our chain-share breakdown reflects what facilitators ADVERTISE, not necessarily where settlements actually HAPPEN. Independent observers measuring on-chain USDC transfers report different breakdowns — Linux Foundation and ETHNews reporting from early 2026 cite Solana with 49–65% of actual settlement traffic. When facilitators publish per-payment streams (rather than aggregate rollups), this dashboard will reflect that finer-grained truth.
Independent observers (Artemis, Allium) report that ~50% of x402 transactions and ~80% of volume are linked to wash trading or self-trading rather than organic agent commerce. Bazaar APIs expose only aggregate counts (l30DaysCalls, l30DaysUniquePayers) — no per-payer distribution — so we cannot detect wash trading directly. The closest proxy we can compute is calls-per-unique-payer: if 99% of a server's calls come from one wallet, that wallet has the same fingerprint as a wash bot. We bucket each server with sufficient sample (≥ 100 calls AND ≥ 5 payers in 30 days) into one of four flags: Verified (< 100 calls/payer), Monitor (100–999), Concentrated (1k–9.9k), Suspect (≥ 10k). Servers below the sample threshold are flagged Unknown. A Suspect flag does not prove wash trading — it means "ask questions before quoting these volume figures." We surface the median across servers and the per-flag distribution on the home dashboard.
Once daily (00:05 UTC) we probe the top 100 indexed servers by 30-day volume with an HTTP HEAD request, 5-second timeout, User-Agent "x402.be-healthcheck/1". A server is recorded as responsive if it returns status 200 (open endpoint) or 402 (canonical x402 challenge). Anything else — other 4xx, 5xx, network error, or timeout — is recorded as unresponsive. The result is exposed publicly at /api/v1/health and the responsive count is shown on the home dashboard. We probe only the top 100 to stay polite and within Vercel function timeout limits; the long tail of registrations is left alone.
A server appears the moment a facilitator indexes it. CDP's Bazaar catalogs services automatically the first time it settles a payment for that endpoint. PayAI works similarly. There is no listing fee, application, or editorial process. Servers can request removal upstream.
Default ranking is descending by 30-day USDC volume. There is no editorial weighting, no paid placement, and no algorithm beyond the sort key the user selects.
x402.be is independent. It is not affiliated with Coinbase, the x402 maintainers, any facilitator, any indexed server, or any other commercial entity. Code is open-source under MIT; data is published as CC0.
The index is a sampling of x402 traffic, not the entire protocol. Servers that route exclusively through facilitators we do not yet aggregate will not appear. Self-reported figures are not independently verified. Volume conversion assumes USDC; non-USDC settlements are valued at the source's reported totalUsd when present, otherwise excluded from volume.