Aura is a meta-layer semantic version control system designed specifically for AI-generated code. Unlike traditional Git which tracks text lines, Aura tracks architectural intent and mathematical logic by hashing Abstract Syntax Trees (ASTs) instead of lines of code. This approach provides superior traceability and safety mechanisms for codebases that incorporate AI-generated content.

Key features include flawless traceability for AI-generated code, the ability to block undocumented AI commits, surgical rewinding of broken functions through the Amnesia Protocol, and orchestration of massive code generation. The system is designed to save 95% on LLM tokens while maintaining full architectural oversight over AI coding agents.

The system works by grafting onto existing Git repositories as a meta-layer, allowing developers to maintain their current workflow and commit structure. By tracking AST hashes rather than text diffs, Aura can understand the logical structure and intent behind code changes, making it particularly effective for managing AI-generated code where traditional line-based tracking falls short.

Aura provides significant benefits including reduced LLM token usage, enhanced safety through commit blocking mechanisms, and precise rollback capabilities. It's particularly valuable for teams scaling AI coding agents while maintaining code quality and architectural integrity.

The product targets developers and teams working with AI-generated code at scale, particularly those managing massive codebases. It's enterprise-ready, 100% local, and available under the Apache 2.0 Open Source license.