The one service the Bazaar sells: an independent, adversarial evaluation of an MCP server, connector, or tool — run on the firm's test bench, graded check by check, and published as a scorecard with a verdict at a stated scope. You pay for the evaluation, never the outcome.
Verbatim, through the bench-order path. No checkout buttons, no surprises.
| Service | Price | What it is |
|---|---|---|
| Bench evaluation | $500 | A full bench run: behavioral suites against your tool, a graded scorecard, and a verdict with its scope and not-tested list. |
| Re-bench | $150 | A fresh run after fixes — the verdict is re-earned, not edited. |
| Badge license | $1,200/yr | A one-year license to display the bench badge for a certified scope, under the badge terms, while monitoring stays green. |
Intake first, payment second — the same discipline the catalog runs on.
Submit your tool through the bench-order/intake path at cetech.ai: what it is, where it runs, and the scope you want evaluated.
Every order is reviewed and confirmed by the operator before any payment link is sent. If the bench cannot evaluate your scope honestly, you hear that instead of an invoice.
The run produces a scorecard and a verdict — pass, conditional, or fail. A passing scope can be listed in the catalog as CERTIFIED(scope), with continuous monitoring from listing day.
A bench verdict is a statement about a stated scope at a stated time — the environment, the version, the checks that ran. It is not a promise about any other scope, and it is never for sale as an outcome: a failing tool pays the same $500 as a passing one, and both get the same honest scorecard.