Real-world Corpus Benchmark

MoonDocKit rendered against 4 real, externally-authored MoonBit documentation sets. Generated 2026-09-13T16:22:50.

4
Projects
4
Rendered OK
0
Failures
100.0%
Pass rate
152
Markdown files
13
Pages rendered
564,684
Source chars

mizchi/markdown.mbt

quality gate passed
Markdown files24
Source characters205,719
Pages rendered5
quality 100/100 205,719 chars 24 md files 5 pages

oboard/mocket

quality gate passed
Markdown files3
Source characters6,441
Pages rendered3
quality 100/100 6,441 chars 3 md files 3 pages

moonbit-community/rabbita

quality gate below threshold
Markdown files38
Source characters79,830
Pages rendered0
quality 75/100 79,830 chars 38 md files 0 pages

mizchi/js.mbt

quality gate passed
Markdown files87
Source characters272,694
Pages rendered5
quality 100/100 272,694 chars 87 md files 5 pages
What this measures. These repositories were never written for MoonDocKit. The benchmark runs the real renderer against their documentation and records compatibility and quality. A "quality gate below threshold" result (e.g. rabbita) is itself a finding: it shows where MoonDocKit's publishing gate blocks low-coverage output rather than shipping broken docs. This is compatibility evidence, not a claim that these projects adopted MoonDocKit.