Arxic
Evidence-driven behavioral intent compiler
Arxic turns a description of what software should do into replayable Playwright workflow bundles — each one carrying provenance and coverage. The point is not to generate tests faster. It is that the result is an artifact you can re-run and cite, rather than a claim that the behaviour was checked.
Why it exists
A passing end-to-end suite tells you a run succeeded. It rarely tells you which behaviours were actually exercised, or lets anyone reproduce that exact run later. When an agent writes the tests as well as the code, that gap stops being academic.
What it does
Replayable bundles
Workflows compile to Playwright bundles that can be re-run, not just reported on.
Provenance
Each bundle carries where it came from, so a result can be traced back to the intent that produced it.
Coverage as output
What the workflow actually exercised is part of the artifact, rather than a separate number to trust.
Where this actually stands
Early. Arxic is being built toward becoming a Ranex feature rather than a standalone product, so the interfaces are still moving.
The repository states its own limits before it states its features. If this page and the README ever disagree, the README is right — tell me and I'll fix the page.
How it fits with Ranex
Arxic is planned to fold into Ranex as a feature. Ranex judges evidence; Arxic produces a particular kind of evidence — behavioural, replayable, and bound to what was actually covered.
What the Ranex kernel enforcesTry it. Break it. Tell me what broke.
Arxic is early, and the useful feedback right now is where it falls over. If you find a hole, that's a contribution.