Live proof
Every chain. Every claim. One page.
We run on more than one chain, so the evidence should not be scattered across more than one page. This is the index: what we registered, what we deployed and what we settled, per chain, each with the transaction behind it. Open a rail to have its facts re-read from the chain while you watch.
Settlement logs
Each rail keeps its own log and is counted on its own, never added to another, because one figure across rails describes none of them. A log keeps a retained window, so its count is the rows it holds rather than an all-time total. Payments from wallets we own are labeled wherever the log can tell.
Settlement logs last read at 2:18:46 PM, 0 of 6 answering. The same 6 logs, read through one shared spec, feed /stats, so the two pages cannot drift apart.
The backend is waking up or briefly unreachable (free tier). Give it a few seconds and refresh.
Published rails
One card per rail. The counts come from our own ledger; the live re-read of every address and token id happens on the rail page itself.
What a caveat count means
Every rail carries a caveat list, and a test fails our build if one is empty. The number on each card is how many things we say are NOT true on that chain: what is testnet, what has no identity registry to anchor to, which payments were between wallets we both own. A proof page that only lists wins is a brochure, so those counts are printed next to the transaction counts rather than under them.