The database is a file stored on Filecoin. Its name is its content: a CID, the hash of the bytes, so anyone holding a copy can prove it is the right one. The IPNI index chain records which storage providers hold that CID, and each provider proves on chain, every proving period, that it still has the data.
Paste the CID. The page reads the index to list the providers, your browser fetches the file straight from the one you pick and checks every block against its hash, then DuckDB, running in this page, runs live SQL on the verified bytes. Load a second provider and the two copies are compared byte for byte. No gateway, no vendor, nothing of ours in the path.
| provider | bytes | blocks verified | sha256 | fetched | match |
|---|---|---|---|---|---|
| – | |||||
| date | transactions | onboarded_pibs | … |
|---|---|---|---|
| – | |||