Problem
There is no public evidence the protocol is alive. Grant reviewers and pilot
partners want traction numbers, and the README itself defines success as
"sustained on-chain attested inference volume" — with no way to see it.
What to do
/status showing service health (API, indexer lag, RPC connectivity) and
protocol metrics: attestations recorded over time, active models and payers,
settlement volume, disputes raised and resolved.
- Sourced from the indexer, cached, degrading gracefully.
- Be honest during the pilot phase: label testnet data clearly and do not
present small numbers as larger ones. A reviewer who catches inflated framing
discounts everything else on the site.
- Publish incident history once there is any.
Acceptance criteria
Notes
Honest small numbers with a visible trend are more persuasive than impressive
numbers a reviewer cannot reconcile with the repository's commit history.
Problem
There is no public evidence the protocol is alive. Grant reviewers and pilot
partners want traction numbers, and the README itself defines success as
"sustained on-chain attested inference volume" — with no way to see it.
What to do
/statusshowing service health (API, indexer lag, RPC connectivity) andprotocol metrics: attestations recorded over time, active models and payers,
settlement volume, disputes raised and resolved.
present small numbers as larger ones. A reviewer who catches inflated framing
discounts everything else on the site.
Acceptance criteria
Notes
Honest small numbers with a visible trend are more persuasive than impressive
numbers a reviewer cannot reconcile with the repository's commit history.