The problem
On a building site the client's inspector writes defect remarks: props too far apart, honeycombing at a column, missing rebar spacers. That log usually sits in the contractor's office or in a database someone else runs. Before acceptance a line can be edited, lost or marked "closed", and nobody can prove what was written.
How SiteLog works
Remarks are written to Arkiv from the inspector's wallet as readonly entities. The page trusts a record only if its on-chain $creator is an inspector in the roster published by the client's wallet. A field anyone can type, like "author", is never used.
See it below
The live journal is read straight from the public Tiramisu RPC: there is no SiteLog server. The demo contractor tried to forge a "resolved" remark and to close its own defect. Both are on chain, and both are shown and ignored.
Another project or trust root
The roster counts only if its on-chain $creator is this client wallet.
Journal verified remarks
Forged and unverified records claim to be remarks for this project, but their creator is not an inspector
Read it without SiteLog
The same data with no SiteLog code: send this query to the public Tiramisu RPC. It is the exact query this page just ran for verified remarks.
Write with your own wallet (MetaMask or any EIP-1193 wallet), Tiramisu testnet
Anyone can write. Writing does not make you trusted: if your wallet is not in the roster, your remark lands under "Forged and unverified" and your closure is ignored. Test GLM: Arkiv faucet.
New remark (inspector)
Claim a fix (contractor)
Close a remark (inspector)
Keep a remark alive (anyone)
Remarks are created with permissionless extension: any wallet can push the expiry later, nobody can bring it earlier.
Live events WebSocket, no polling
- Waiting for Arkiv entity events on this project…