How to read x402 dashboards without being fooled

Three totals, three scopes

Coinbase published its own cumulative count in late April 2026: roughly 165 million x402 transactions, about $50 million of volume and 69,000 active agents, in the same week Agent.market launched, as CoinDesk reported via Yahoo Finance. Chainalysis counted well over 100 million cumulative x402 transactions on Base through Q1 2026. TRM Labs counted 198.9 million settlements worth $52.7 million across Base, Solana and Polygon since May 2025.

Three numbers, no two of them measuring the same thing. A reader quoting 165 million and a reader quoting 198.9 million are both citing a primary source. Neither is describing the market.

Read the table by column, not by row. The headline figure is the least informative cell in it.

Chain scope is the first thing to check

TRM Labs counts three chains. Coinbase's number is Base. Chainalysis's 2026 count is Base. Add Solana and Polygon and the total grows without a single new payment appearing anywhere.

The window matters just as much. TRM starts in May 2025 and runs to 9 September 2026. Chainalysis stops at Q1 2026. Coinbase's figure is a snapshot from late April 2026. A longer window on more chains will beat a shorter window on one chain almost every time, and that tells you nothing about which network is healthier.

Then ask what a "transaction" is. TRM found 99.6% of value settled in USDC, the stablecoin documented by Circle's USD Coin (USDC) product page. A dashboard that aggregates transfers across many assets can show a rising count while the value that matters sits in one stablecoin.

Facilitator and indexer scope

An x402 payment is verified and settled by a facilitator, as Coinbase Developer Platform's x402 launch announcement states. Whichever facilitator's traffic an indexer watches becomes the indexer's universe. Two dashboards can both be accurate and still disagree by a factor of two because one sees a single large facilitator and the other sees the long tail. If you want the background on that layer, what is an x402 facilitator, and do you need one? covers where the settlement work actually happens.

x402scan is the third-party indexer most people open first. It tracks transactions, sellers, origins and resources across Base and Solana. Treat it as a live instrument: cite it for what it displays and for how to interpret it, never for a fixed total you paste into a document, because the number will have moved by the time anyone clicks your link.

There is no protocol-level ledger to reconcile against. According to the x402 documentation, the protocol charges no fee of its own and payers cover network fees only, across EVM chains and Solana. Every published total is somebody's view of somebody's traffic.

What no dashboard can tell you

No dashboard fields a "was this real" column. TRM's $52.7 million screens down to $25.62 million of plausible commerce, and between 0.6% and 7.5% of that screened commerce looks agentic. That is a judgement call about which payments look like commerce, made after the fact by an analyst.

Volume and average size can mislead in the other direction. CoinDesk, citing Artemis, put real daily x402 volume at around $28,000 with an average payment around $0.20 against roughly 131,000 daily transactions (CoinDesk). Do the arithmetic and it roughly holds. It also describes a market that is tiny, which is the point.

Cumulative counters hide direction. BlockEden.xyz reported daily x402 transactions falling from about 731,000 in December 2025 to about 57,000 in February 2026, a decline of over 92%. Coinbase's cumulative figure kept climbing through the same period, because a cumulative line cannot go down. Meanwhile Chainalysis found payments of $1 or more rising from 49% of value in early 2025 to 95% by early 2026, so the mix changed while the totals rose.

None of that is visible on a chart that only plots a running sum.

Questions to ask before you quote a figure

1. Which chains are inside the count, and which are outside?
2. Whose traffic does the source see: one facilitator, several, or a crawl of the chains?
3. Is it cumulative or daily, and what is the closing date?
4. Has anything been screened out, and by what rule?
5. Is the agentic share measured, or inferred from payment size?

A figure that survives all five is quotable. A figure that fails one is still usable, provided you say which one it fails.

Why the disagreement is useful

The spread between Coinbase, Chainalysis and TRM is a map of scope, not a scandal. Coinbase counts what its own rail processed on one chain. Chainalysis counts Base through a fixed quarter. TRM counts three chains over the longest window and then removes what does not look like commerce.

If you are pricing an API for agents, the screened number is the one that should set your expectations. How to price an API for AI agents walks through that, and the same discipline applies to reporting: state the scope before the number. When a vendor quotes a headline total at you, ask for the four columns in the table above. If they cannot produce them, you have your answer.

This is the same problem as AEO compared with SEO: a metric only means something once you know who measured it, over what, and why. The method behind every figure published here is public for exactly that reason.