Aura: Agents + Git + Intent IDE

Product Introduction

1. Definition: Aura is a Git-native Integrated Development Environment (IDE) and control layer specifically designed for managing AI coding agents. It is a semantic overlay that integrates directly with local Git repositories to provide AST-level tracking, intent logging, and autonomous task management for AI-assisted development.
2. Core Value Proposition: Aura exists to solve the control and provenance problem in AI-powered software development. While AI agents can generate code in seconds, Aura ensures developers can understand, verify, and own the changes. Its core proposition is providing semantic version control for AI agents, enabling provenance tracking, intent-aware diffs, and proof-ledger commits on top of standard Git workflows.

Main Features

1. Mission Control (Multi-Agent Orchestration): A unified desktop interface that allows developers to run and monitor multiple AI coding agents—including Claude Code, Gemini CLI, Antigravity, Cursor Agents, and Codex—simultaneously within a single project window. How it works: It acts as a central MCP (Model Context Protocol) server, allowing different agents to connect natively. All agents share one conversation history and composer, enabling context switching without loss and providing a consolidated view of all agent activity.
2. Semantic Engine & AST-Level Diff: The foundational technology powering Aura is a Rust-based semantic engine that analyzes code at the Abstract Syntax Tree (AST) level, not just line-by-line. How it works: It creates a mathematical identity for functions and classes, making diffs rename-proof and logic-aware. This allows developers to see what changed in the logic (e.g., a modified conditional statement) under a "why it changed" header, rather than just seeing red/green line changes.
3. Crew (Autonomous Task Work-Loop): An autonomous system that manages a backlog of coding tasks. How it works: Developers provide a stack of tasks with dependencies. Crew agents claim unblocked tasks, work on them in isolated Git worktrees (eliminating branch juggling), and each completion is automatically verified against its original goal before being considered for merge. This enables unattended, parallel agent execution with built-in validation.
4. Trace & Proof Ledger: A cryptographically sealed provenance system. How it works: Every code change (commit) is permanently linked to the AI agent session that produced it, including the full conversation transcript and the specific intent or goal. This creates a verifiable "genuine-record" that ties the commit to the goal it delivered, providing an audit trail for who (which human/agent) wrote what and why.
5. Git-Native Integration: Aura does not replace Git but grafts onto it. How it works: It uses Git hooks and Git notes to attach its semantic metadata and intent records directly to the repository. All standard Git commands (commit, push, rebase) remain functional. The provenance data is stored on a dedicated orphan shadow branch to survive operations like rebase.

Problems Solved

1. Pain Point: Loss of Control and Understanding in AI-Driven Development. Traditional tools show line-based diffs, which fail to convey the semantic intent behind AI-generated changes. Developers struggle to verify if an AI agent actually completed the intended task correctly.
2. Pain Point: Agent Collision and Context Fragmentation. Using multiple AI agents (e.g., Claude in Cursor, Gemini in CLI) leads to scattered context across different tabs and tools, making it hard to maintain a coherent project history and causing parallel agents to overwrite each other's work.
3. Target Audience: Engineering Teams and Lead Developers adopting AI coding assistants at scale; Open-Source Maintainers reviewing AI-generated pull requests; Tech Leads and Engineering Managers who need audit trails and quality assurance for AI-assisted work; Individual Developers who use multiple AI tools and want to maintain code ownership and understanding.
4. Use Cases: Automated Code Refactoring with verifiable safety; Autonomous Bug Fixing where Crew works through a backlog of issues; AI-Paired Programming Sessions with full transcript logging for later review; Code Review and Audit, where a reviewer can inspect the sealed intent and proof ledger for any commit.

Unique Advantages

1. Differentiation: Unlike AI-powered IDEs (Cursor, Cline) or standalone CLI agents, Aura is a control plane and provenance layer that works across all of them. It provides a unified semantic history and proof system that other tools lack. Compared to traditional Git GUIs, it offers AST-level semantic diffs and intent logging, not just line-based changes.
2. Key Innovation: The "Goal ↔ Commit Proof Ledger" is a fundamental innovation. It moves version control from tracking "what changed" to tracking "what changed and did it fulfill the promised goal?" This is enabled by the rename-proof AST-level identity and the cryptographically sealed session records that link agent reasoning directly to code changes.

Frequently Asked Questions (FAQ)

1. Does Aura replace my existing Git workflow? No, Aura is a meta-layer that integrates with your existing local Git repository. You continue to use git commit, git push, and manage branches as usual. Aura enhances Git by adding semantic diffs, intent logging, and provenance tracking via Git hooks and notes, allowing you to revert to plain Git at any time.
2. How does Aura handle my code privacy and data security? Aura operates fully locally. Your code, Git history, and AI session transcripts never leave your machine unless you explicitly push to a remote repository. The system captures provenance for verification but does not harvest or train on your data.
3. Can I use Aura with the AI coding tools I already have? Yes. Aura connects natively via MCP to agents like Claude Code and Gemini CLI. For AI-powered editors like Cursor or Aider, it automatically captures transcripts and intent via hooks. It is designed as a unified platform for the AI agents your team already uses.
4. What happens to Aura's tracking data during a Git rebase or merge? The data is preserved. Aura stores checkpoint and provenance data on a dedicated orphan shadow branch with sharded storage. When commit hashes change during rebase, stash, or merge, Aura detects the HEAD migration and re-anchors all semantic history, ensuring traceability survives standard Git operations.
5. Is Aura suitable for team collaboration? Yes. Features like the Team Radar provide a live awareness plane showing who (or which agent) is editing what, preventing collisions. The sealed, portable genuine-records allow any team member or reviewer to verify the provenance and intent behind any change without needing the original author's context.