Hey HN,

  I built DevOps Agents — a set of specialized AI agents to help with day-to-day DevOps and SRE work.

  The agents analyze your GitHub repo, figure out the cloud resources needed, deploy everything, and get your app live in
  production. They have a chat interface (like Claude Code or ChatGPT) and stick around after deployment to help you manage the
  infra ongoing. For example, I can ask the agent if the app is running, and it will SSH into the resource, check the status,
  check logs, find the root cause, and get the app back up.

  I've used this for various tasks involving complex setups — Kubernetes, ELK stack, Grafana, Prometheus, Redis, ClickHouse, CI/CD
   pipelines, self-hosted tools, Docker setups, and multi-account AWS, Azure, GCP, or DigitalOcean deployments.

  Why I built this: I was already using Cursor and Claude Code to manage my infra, and they're great for code. But infrastructure
  is different — most of the context isn't in code. Things like where your configs live, which cloud accounts map to which
  environments, what ports are exposed, how services talk to each other — all of that lives outside the codebase. Every time I
  started a new session, I was re-explaining my entire setup from scratch. The agents here are purpose-built to retain that
  infrastructure context across sessions, so I can keep managing my infra without losing continuity.

  Happy to answer any questions.