Aura's maintained source and documentation are in youngbryan97/aura. This older repository contains a landing page rather than the current engine implementation.
Aura is a local AI research project built around a language model, lasting records, attention, action control and ongoing internal state. The practical goal is an assistant that can keep track of work, use relevant history, explain its actions and remain responsive while it works.
- The plain-English tour explains what the parts do, why they matter and how they apply in ordinary use.
- Variables and examples translates the numbers and technical terms, with examples and metaphors.
- The full documentation map separates current guides from dated evidence, standards and generated pages.
- Demos and related work connects mechanisms to evidence and primary research without treating a planned demonstration as a witnessed result.
- Scientific foundations and sources explains what the cited research supports, how Aura adapts it and which results need Aura's own evidence.
ISC explained introduces the experiments. The completed seed-23 run passed 22 of 24 operational checks under ISC-v6, with positive content and withheld-report outputs. A later isolation defect means those paired findings need rechecking. The carrier and split-system control questions remain unsettled. Read current status for the exact measurements and later developments.
The separate proof-kernel repository is a small demonstration of earlier components. Its name does not describe a completed proof of consciousness, and this landing page supplies no additional experimental result.