The AstroPulse Journey
When I started AstroPulse, the problem was easy to name and hard to live with: teams moving to the cloud were drowning in tools. Every provider had its own consoles, its own primitives, its own way to deploy an app and stand up a cluster. The work that mattered, shipping software, kept getting buried under the work of operating it.
I wanted one place to deploy an application and run a cluster, on any cloud, without learning five different platforms first. That was the beginning. Everything since has been built on top of that one idea, one layer at a time.
The foundation: clusters and apps, on any cloud
We started at the bottom of that picture. Astro Platform gave teams one way to provision and operate Kubernetes, with astroctl as the single interface across AWS, GCP, and Azure, plus self-hosted clusters and existing clusters brought in through registration. Deploy an app from an image, a Helm chart, a repository, or plain YAML. Pull a kubeconfig for any cluster on any cloud with one command.
It was unglamorous and it was the right place to start. You cannot earn trust to operate production until you can reliably reach production in the first place.
The turn: the control layer was the point
The lesson came from watching how teams actually used it. A model with connectors can tell you what broke. Describing a problem is becoming a commodity. The hard part, the part everyone underestimates, is the trust to act in production: a change that runs inside a scope it cannot exceed, with policy guardrails, human approval where it matters, a full audit trail, and a way back.
So we built that layer in the middle of the picture. The model was never the bottleneck. The control layer was. Once that existed, the same governed runtime could serve a person typing a command and an AI proposing one. One set of guardrails, not two.
Where we are today: two products, one runtime
That is where we are now, at the top of the picture. AstroPulse is two products that work together. Nova is your AI platform engineer: it investigates, plans, and acts across your stack, and incident response, the AI SRE work, is one of the things it does especially well. Astro Platform is the governed runtime Nova goes deepest on, and the operational control plane your team uses directly through the Console and the CLI.
The interesting part is not that an AI can suggest a change. It is that it can make one safely, under governance you control.
Where we are going: beyond Kubernetes
If you read our very first post, the direction was already there: the cluster abstraction was always meant to evolve. Kubernetes is the substrate we operate today, not the boundary of the idea. The control layer is the durable part of the system, and we intend to extend that same governed-action model to more of your infrastructure over time. Autonomy is earned as patterns prove safe, never assumed.
Thank you
We are here because people trusted us early, before any of this was finished. To our pilot customers who ran real workloads on a young platform and told us the truth about what was missing, and to the developers and the broader community who filed the issues, asked the hard questions, and pushed us to do this properly: thank you. You shaped every layer in that picture.
The foundation is solid, the control layer is real, and the most exciting work is still ahead. Things are going to be great going forward, and we are glad to be building it with you.
Meet Nova, your AI platform engineer: Open Nova. Or read the platform overview to see how the layers fit together.