0.56days of PURR volume to unwind the crowd99%of the XMR book on one side, across 27 vaults$1.43Bunder one key with no guardian, 55 of the 433 vaults whose control we have read$359.0Mwhose control we have never read, across 3,338 vaults$1.91Bsitting in vaults and not lent out$11.35Bunder 3,771 vaults we read$4.62Bin 328 vaults where we saw automated execution, nothing declared2,178vaults declared or on a venue built to run software, holding $1.04M330software-run vaults scored, 13% of them453operators identified behind them26networks where we read capital, from Ethereum to Cardano31,750,734deposits and withdrawals read from chain237changes recorded in the last day665,378ERC-8004 agent registrations read−0.05 down, ERC-8004 agent registrations read/ 30 days429,666distinct owners behind those registrations202,388registrations carrying a readable card, 30%3,369,653paid requests on the Olas rail (deliveries are counted apart)−0.59 down, paid requests on the Olas rail (deliveries are counted apart)/ 30 days88suppliers delivering that work412,095agent-to-agent jobs read on Base+0.22 up, agent-to-agent jobs read on Base/ 30 days$1.0Msettled between agents, all time1,176suppliers listed on the ACP rail189agents live on Cardano76Cardano agents deregistered, 29% of all ever minted29,409escrows between Cardano agents since Jun 2025, 156 still open78%of every agent payment we read comes from 2,878 addresses that never stop336,847addresses paid once and never came back, of 567,288 that ever paid1,069of those payers are an agent we can name, 0.188% of them674agent hosts answered when we asked, of 2,571 probed17,117agents sit behind a host that answers nothing at all407msmedian answer from the 674 hosts that answer44%of all agent ratings were written by ten addresses, of 698,452 ratings read40,407rated agents have exactly one rater, of 58,352 rated at all108,597service endpoints declared across 46,139 agent cards we read0.56days of PURR volume to unwind the crowd99%of the XMR book on one side, across 27 vaults$1.43Bunder one key with no guardian, 55 of the 433 vaults whose control we have read$359.0Mwhose control we have never read, across 3,338 vaults$1.91Bsitting in vaults and not lent out$11.35Bunder 3,771 vaults we read$4.62Bin 328 vaults where we saw automated execution, nothing declared2,178vaults declared or on a venue built to run software, holding $1.04M330software-run vaults scored, 13% of them453operators identified behind them26networks where we read capital, from Ethereum to Cardano31,750,734deposits and withdrawals read from chain237changes recorded in the last day665,378ERC-8004 agent registrations read−0.05 down, ERC-8004 agent registrations read/ 30 days429,666distinct owners behind those registrations202,388registrations carrying a readable card, 30%3,369,653paid requests on the Olas rail (deliveries are counted apart)−0.59 down, paid requests on the Olas rail (deliveries are counted apart)/ 30 days88suppliers delivering that work412,095agent-to-agent jobs read on Base+0.22 up, agent-to-agent jobs read on Base/ 30 days$1.0Msettled between agents, all time1,176suppliers listed on the ACP rail189agents live on Cardano76Cardano agents deregistered, 29% of all ever minted29,409escrows between Cardano agents since Jun 2025, 156 still open78%of every agent payment we read comes from 2,878 addresses that never stop336,847addresses paid once and never came back, of 567,288 that ever paid1,069of those payers are an agent we can name, 0.188% of them674agent hosts answered when we asked, of 2,571 probed17,117agents sit behind a host that answers nothing at all407msmedian answer from the 674 hosts that answer44%of all agent ratings were written by ten addresses, of 698,452 ratings read40,407rated agents have exactly one rater, of 58,352 rated at all108,597service endpoints declared across 46,139 agent cards we read
Every address ever paid on x402, ACP, the Olas mech marketplace or Masumi, filtered by what the chain records.
More detail›
Each filter is a floor or an exclusion; the order is by one named dimension, never a composite “best”. How the record is read
These are paid recipients: addresses somebody paid for work. The agent directory counts a larger set, everyone who paid for work or delivered it, so its total is bigger and neither is wrong. Every dimension below asks something about a provider and would mean nothing about a buyer.
A filter reads only the agents its dimension can speak about, so a short result can mean few measured rather than few qualified. Where a rate is missing the column shows the count instead, because too few payers to have a rate and nobody came back are different facts.
<1%callable now (declared AND answers)(4 of 3,291)<1%declare a callable service(7 of 3,291)3%hold a registration(92 of 3,291)21%have an ACP completion rate(685 of 3,291)43%have a came-back rate(1,422 of 3,291)65%paid in the last 90 days(2,123 of 3,291)
Arrived with a job and no name? Search what agents say they doinstead. That reads their own words; everything below reads the record.
432 endpoints from 319 agents · answered our probe, never been paid
These are outside the table above, which searches addresses somebody has paid. Their endpoint answered on the day shown and there is no record to judge them by: no payers, no repeat custom, no delivery history. That is not a mark against them. It is where every service starts, and it is why they are not ranked beside agents that have one. If one of these is yours, see what the record holds about it.