updated 2026-09-14 · on-chain data complete since 2025-05-09
The State of x402
A live, data-driven map of the entire x402 payment ecosystem.
We continuously probe every listed x402 endpoint and watch every USDC
settlement on-chain (Base + Solana). The headline: of 137,878 listed endpoints,
55% are unreachable, 5% of the reachable ones serve a
non-compliant payment envelope, and only 59.9% of the pay-to
wallets behind them have ever received a payment. The directory is mostly ghosts.
The ecosystem right now
82,051,564 probes and counting45%
actually reachableof 98,456 active endpoints
95%
spec-compliant 402of the reachable ones
60%
wallets ever paid3,759 earning wallets
137,878
endpoints listed
98,456
currently active
39,422
delisted
82,051,564
probes executed
Real money settled (on-chain)
$56.08M
total volume
229,830,420
settlements
175,099
paying wallets (tracked)
3,759
wallets that ever earned
Total volume is measured across all known x402 facilitators: every USDC settlement one of them submitted on-chain (Base + Solana). Coverage depth: Base is complete since May 2025; Solana history starts at each facilitator address's registry date (deep-history windows for the largest Solana facilitators were backfilled 2026-08-26; long-inactive tail addresses may start later). Per-facilitator market share: the facilitator leaderboard. Per-wallet figures below (top earners, paying wallets) cover the 3,759 distinct pay-to wallets we resolve to a listed endpoint; one seller often runs many endpoints off a single wallet.
Monthly momentum
▲ +62% 2026-08 vs 2026-07
Where the money goes (top earners)
click a column to sort| # | endpoint▼ | volume▼ | share▼ | settlements▼ | payers▼ |
|---|---|---|---|---|---|
| 1 | api.bitrefill.com | $492K | 27.5% | 4,091 | 1,688 |
| 2 | 3MZVk9…NfYX (unlisted) | $446K | 24.9% | 1,232 | 433 |
| 3 | swarmx.io | $177K | 9.9% | 10,927 | 14 |
| 4 | mcp.x402.boats | $78K | 4.3% | 18,921 | 35 |
| 5 | api.surplusintelligence.ai | $72K | 4% | 15,628 | 95 |
| 6 | blockrun.ai | $63K | 3.5% | 4,561,271 | 388 |
| 7 | api.clusterprotocol.ai | $61K | 3.4% | 64,015 | 721 |
| 8 | x402endpoint-utiwoa54hq-uc.a.run.app | $53K | 2.9% | 661 | 93 |
| 9 | FPfGD3…7Vk4 (unlisted) | $50K | 2.8% | 123 | 42 |
| 10 | AQqnMF…EeKA (unlisted) | $30K | 1.7% | 1,026,318 | 68 |
What's broken (402 envelope issues, 30d)
no-accepts9,297,961missing-x402-version9,149,361http-status-4045,342,932no-envelope3,350,764http-status-5031,534,053http-status-2001,127,166http-status-502833,646http-status-400730,643missing-resource-info419,110http-status-530237,060HTTP-native vs MCP-native
reliability split by probe transport
HTTP-nativeplain HTTP 402 paywall
98,327
active endpoints · 99.9%
53.7%
avg uptime (scored)
A · 12,535 B · 31,085 C · 7,706 D · 3,886 F · 35,976 ? · 7,139
MCP-nativex402 challenge served over MCP
129
active endpoints · 0.1%
78.2%
avg uptime (scored)
A · 0 B · 55 C · 50 D · 7 F · 17 ? · 0
MCP-native services never answer a bare HTTP request with a 402: the challenge only appears after an MCP handshake (
initialize → tools/list → an unpaid tools/call). We probe both paths without ever paying and grade the payment envelope identically, so the two populations are directly comparable. MCP probe latency includes the handshake and is not 1:1 comparable to HTTP latency.