# Architecture Diagram Reviewer

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Sathishkannan Sakthivel](https://cerebralvalley.ai/u/sathishkannan)
- **GitHub:** https://github.com/sathishkannan-labs/Architecture-Diagram-Reviewer
- **Demo video:** https://youtu.be/m2SEJ6It16g?si=o8UMgvEB8XSBaKBt
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4503

Architecture Diagram Reviewer is an AI-assisted solution that transforms architecture diagrams into evidence-based technical reviews and traceable Infrastructure as Code.

Using Meta's multimodal AI capabilities, the application extracts components, connections, and architecture boundaries from uploaded diagrams. Users can inspect and correct the extracted graph before initiating a structured risk review.

The review distinguishes confirmed findings from unverified questions, helping architects identify issues such as single points of failure, database exposure, missing failover, and unclear security controls without treating assumptions as facts.

Users can then review and approve a supported AWS target architecture. The application generates deterministic Terraform from trusted templates, with resource-to-decision traceability and a separate offline validation workflow.

Built with Python, Streamlit, FastAPI, Meta AI, and Terraform, the project combines AI-assisted understanding with human oversight, explicit approval, and auditable infrastructure generation.

The goal is to make architecture reviews faster, more explainable, and actionable without giving AI unrestricted control over infrastructure code.

Live Demo Access

Application: https://adr.sathishkannan.com

Access Code: c6274da9900a63717527737bba9ba7c2

Judges can upload their own architecture diagrams (PNG/JPEG) to explore AI-powered extraction, architecture risk analysis, human-in-the-loop review, and traceable Terraform generation for supported AWS architectures.

## More from Meta Global AI Developer Hackathon

- [TWZRD](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4500)
- [Museverse](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4501)
- [MoreAI](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4502)
- [Clarifycue](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4504)
- [MAKERS](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4505)
- [PackProof](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4506)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4503. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt