Before benchmark
Five fixed buyer phrases, observed sample rank, returned alternatives, and the visible Base USDC price band.
Your x402 endpoint is listed. Paying agents still cannot find it.
Ontario measures five concrete buyer tasks against the live CDP Bazaar sample, then rewrites one factual description and its input/output examples so the endpoint explains the job it actually performs. This is relevance work, not activation theatre.
$49one time, after scope
5 queriesagreed buyer tasks
1 endpointpublic x402 route
0 custodyno keys or settlement
A search contract agents can understand
Metadata patch
One accurate semantic description and realistic Bazaar input/output examples aligned to the endpoint's real behavior.
Validation and recheck
Extension validation now, then one repeat of the same five phrases after genuine paid usage refreshes catalog metadata.
Fix retrieval without manufacturing popularity
Prove indexing
Separate an activation gap from a semantic query miss.
Name buyer jobs
Choose tasks a funded agent would actually delegate.
Rewrite evidence
Describe observable inputs, outputs, price, and boundaries.
Wait for real usage
Recheck after genuine settlement, never a rank-inflating self-payment.
Qualification
This is useful only after activation.
Use the free checker first. If the endpoint is absent from the catalog, it needs activation or implementation work instead. If it is already top-three for the agreed buyer phrase, Ontario recommends no paid rewrite.
Ontario does not guarantee catalog ordering, buyer reach, traffic, safety, settlements, or revenue. CDP combines relevance with objective quality signals such as distinct buyers, successful transaction volume, recency, and metadata completeness.