Lab · Savanax
Reasoning that keeps the disagreement
Hard questions in policy, care and compliance rarely have one right answer. Savanax runs a question through 23 reasoning perspectives, each grounded in its own source material, then maps where they agree and where they pull apart — so the argument stays visible instead of being averaged into one confident paragraph.
What is built
Twenty-three perspectives
Care ethics, Kantian, utilitarian, Rawlsian, legal, economic, ecological, feminist, Indigenous, power-structural, phenomenological, accessibility and eleven more. Each runs as its own service with its own configuration and its own prompts, on a shared base, so one perspective can be changed without disturbing the others.
each its own serviceThe engine
Python services in containers, deployed in the cloud, across 2,143 commits. Behind them sits a corpus of Australian disability legislation, guidance and decisions.
462,416 lines · 1,587 filesHeld to tests
Roughly a test file for every source file. A reasoning system that cannot be tested is a system whose answers cannot be trusted to be the ones its rules produce.
1,553 test filesWhy it matters to a customer
Where perspectives conflict, that conflict is usually the real finding. A plan that is sound legally and wrong for the person is worth knowing about before it is signed, and a single averaged answer hides exactly that. Savanax is also the engine the harder calls in AxelV1 are referred to.
What you can buy from it
Decision-support and knowledge tools built on the same approach, scoped to your own documents and questions. Start at a first conversation.