Don't trust us. Verify us.

Every verdict on this page is computed by your browser — hashes of bytes you generated, answers you recomputed, events you caused. The server's claims are never taken at face value. A fake backend cannot pass these proofs.

0 node(s) online · no proofs run yet

Proof 1 — Liveness

not run

Every node on the dashboard is a live device. You can prove it by causing a change yourself.

Proof 2 — Storage integrity

not run

Bytes you store come back bit-identical. Judged by hashing in YOUR browser, before and after.

Proof 3 — Replicated possession

not run

Every replica node physically holds your bytes. Salted challenge — a fake node cannot answer.

Proof 4 — Survives node death

not run

Kill a node yourself; your data still reads back hash-identical from the survivor.

Proof 5 — Honest compute

not run

The grid's job answers match your own browser recomputing the same problem.

Proof 6 — No overselling

not run

The Reserve refuses to sell capacity that does not physically exist online.

Don't even trust this page

This page is served by the same site it verifies. If that bothers you (it should), go around it:

  • Raw state, no UI: curl localhost:3000/api/v1/grid/nodes — the JSON the dashboards are drawn from. Watch it change as you open and close /join tabs.
  • Watch the bytes move: on a /join tab, open DevTools → Network → WS → the /ws connection. When you store data in the console, you will see 4 MB binary frames arriving in that tab. That is your data, entering that device.
  • Task Manager: the RAM a /join tab pledges is allocated up front — watch the tab's memory jump when you click Contribute.
  • Read the source: the entire system is on your disk. There is no server you can't see. server/coordinator.ts is the whole truth.
  • The killer argument: you are running the coordinator yourself, on your own machine. In this deployment there is no third party left to distrust.