Aura AI
Link to open source: https://github.com/Saanvi-gug/AURA https://drive.google.com/file/d/1SMVDl0iBWAOq7FP1Lb4DMpNo3RFpB6u5/view?usp=drivesdk
AURA AI is a next-generation engineering architect designed by team Skygliders to transform unstructured founder visions into technically grounded, audit-ready Business Requirements Documents (BRDs).
At the core of AURA is a sophisticated Multi-Agent Swarm logic that bypasses the limitations of traditional AI by deploying three specialized intelligence layers in parallel:
-
The HAWK Agent (The Grounder): Acts as the technical anchor, mapping high-level intents directly to complex datasets, such as the Olist v2.1 logistics schema, to ensure every requirement is rooted in real-world infrastructure.
-
The GURU Agent (The Risk Auditor): Operates as the system's "cynic," specifically hunting for "Ghost Risks," technical debt, and architectural bottlenecks that general-purpose models often overlook.
-
The INTEGRA Agent (The Architect): Serves as the lead synthesizer, ingesting the outputs from Hawk and Guru to assemble a professional, tabular BRD complete with Executive Summaries and Risk Mitigation protocols.
By leveraging Google Cloud Vertex AI, AURA provides enterprise-grade reliability and precision, ensuring that the final output isn't just a document, but a production-ready blueprint validated by an authenticated agentic workflow.
This build was uploaded as a hackathon project






