Deployments in production
The claims on this site are testable, because one deployment of the engine is public. It belongs to a real standards organization, serves its real corpus, and runs the same gates on every change that your deployment would run.
OIML SMART AI
The public question-answering service of the International Organization of Legal Metrology, running the Konneal engine in production — cited answers, computed verdicts, and the full measurement discipline, purely on open-weight models.
The engine behind it — retrieval over clause-structured corpora, the answer contract, conformance by execution, the evaluation gates — is the same package every deployment consumes from the npm registry, promoted through the same golden suites that grade the live service.
- publications
- ≈ 900
- words indexed
- 9.3 M
- languages
- 9
- defined terms
- ≈ 9 000
- bibliographic records
- 5 707
- citation-graph edges
- ≈ 300
- publication families
- 224
- golden suite
- 95 %+
White-labeled to your organization
A deployment carries your publications, your terminology, your profile and your frontend. Konneal appears only in a small credit beside your emblem.
The credit lockup is specified in the brand: “Powered by” set in the regular weight at reduced ink, the mark at 1.4 × the wordmark’s x-height, baseline-aligned. It sits beside institutional emblems, not above them.
Adoption follows what your organization already has: a Metanorma library is enough to begin, and every further asset raises the annealment your members get.
The adoption paths → Get started