Hey Product Hunt 👋 ArchGenie is an AI Cloud Architect for AWS, Azure, and GCP. You can describe your system in natural language, sketch it on a whiteboard, or upload a screenshot of an existing architecture — and get back a complete cloud solution in one shot: 🏗 Architecture diagram with the right services for your domain 💻 Working Terraform, Pulumi (TypeScript or Python), or Bicep code 🛡 Security posture grade (A–F) — CIS, NIST, HIPAA aligned, with category breakdown 💰 Monthly cost estimate including the long tail — NAT, data transfer, and GPU/TPU pricing for AI/ML workloads (H100, H200, Inferentia, Trainium, TPU v5e/v7) 📊 Observability code — CloudWatch, Cloud Monitoring, Application Insights, plus Grafana dashboards on demand 📚 README that explains what the architecture does and how to deploy it Free tier, no credit card. **Why we built it:** Every cloud project starts the same way — someone draws boxes in a diagramming tool, someone else translates those boxes into Terraform, a third person hardens it for security, a fourth fights with the cost calculator. By the time it's real, the diagram is already stale. ArchGenie collapses that whole loop into a single description-to-deployable flow, and keeps the diagram, code, security posture, costs, and docs in sync as requirements change. **What works well today:** • Three input modes — natural language, whiteboard sketch, or screenshot of an existing architecture • Three infrastructure-code formats — Terraform, Pulumi (TypeScript or Python), and Bicep — each reads like hand-written code in that format • Three providers — AWS, Azure, GCP — with full service catalogs • One generation produces the complete picture: diagram, working code, security grade, cost breakdown, observability, and a README • Cost estimates cover the long tail that usually surprises teams — NAT gateways, data transfer, GPU/TPU SKUs for AI/ML workloads • Iterate in plain English — change the architecture by describing the change; everything stays in sync • Integration and xports for GitHub, GitLab, Bitbucket, and Jira **Where we're still rough:** • Brownfield import isn't there yet — you can describe a new system, sketch one, or upload a screenshot of an existing diagram, but you can't point ArchGenie at a live cloud account and have it reconstruct what's already deployed. Greenfield is fully covered; live-account import is on the roadmap. • One provider per architecture today — AWS-only, Azure-only, or GCP-only designs are first-class. Cross-cloud setups (active/active across two providers, hybrid on-prem + cloud) still need manual stitching. • Deploy pipeline is yours to wire — ArchGenie ships the architecture and everything around it, but the CI/CD layer (GitHub Actions, GitLab CI, Atlantis, Terraform Cloud) is something you bring for now. **Would love feedback on:** 1. Where does the generated infrastructure code fall short of what you'd write by hand? 2. Which integrations matter most for your team — beyond GitHub, GitLab, Bitbucket, and Jira? 3. Which compliance frameworks beyond CIS / NIST / HIPAA would unlock real use cases for you (SOC 2, PCI-DSS, FedRAMP, ISO 27001)? 4. Which cloud providers beyond AWS, Azure, and GCP would your team need supported? (Oracle Cloud, Alibaba Cloud, IBM Cloud, DigitalOcean, Cloudflare, OVHcloud — or something else?) 5. Which infrastructure-code formats or languages would help ArchGenie fit into your existing stack better? (CloudFormation, AWS CDK, ARM templates, Pulumi in Java / C# / Go, Terraform CDK — or something else we're missing?) Try it: archgenie.io — happy to answer questions in this thread.

ArchGenie

AI Cloud Architect — describe it, ship the whole stack

AI Cloud Architect — describe it, ship the whole stack