Cloud CI and agentic workflows for embedded hardware development
Kubernetes operators for embedded hardware labs finally solve the 'works on my bench' problem.
Heroku-like PaaS CLI for vanilla Kubernetes. Zero server-side components.
Provides the familiar Heroku commands (apps:create, deploy, config:set, services:logs) while creating native K8s resources under the hood and even auto-provisioning Ingress + TLS via cert-manager — all without installing an operator. Practical and pragmatic: great as an MVP for teams tired of YAML, but you'll want to vet RBAC, multi-tenant safety and edge-case behavior (builds, scaling, secrets, quotas) before trusting it for production workloads.
Backend developers, DevOps/platform engineers, small teams who want Heroku-like DX on their own Kubernetes clusters
I built Kuberoku, a CLI that aims to bring a Heroku style workflow to vanilla Kubernetes.
I’ve used Heroku for years and still think the apps / deploy / logs / run DX is hard to beat. On the other hand, I like Kubernetes, but I don’t love how much of the day-to-day workflow becomes YAML and glue.
Right now, Kuberoku is still early, but kuberoku deploy works and handles the basic deployment workflow (build + apply + rollout) via Kubernetes primitives.
How it works: Kuberoku is written in Python and interacts directly with the K8s API. It doesn't require a heavy server-side operator; it uses native Kubernetes primitives to handle the build-apply-rollout loop.
I’m currently working on: - a git push-style workflow - RBAC / multi-tenant safety so it’s usable by teams
I’d love feedback from K8s/platform folks on: What’s the first edge case you’d try to break in a tool like this?
Kubernetes operators for embedded hardware labs finally solve the 'works on my bench' problem.
Kubernetes-native AI agent orchestration via YAML; early but self-hosting Kelos development proves the loop.
Declarative identity management via Kubernetes manifests is genuinely useful for Kanidm users.
flarepilot removes the usual Wrangler/Worker boilerplate and lets you build, push and expose a Docker image in one command while auto-creating a routing Worker, Durable Object bindings and DNS. The multi-region routing, live logs, and interactive domain picker are practical touches; it's a pragmatic, opinionated wrapper around Cloudflare's beta Containers that saves manual setup, though its value depends on committing to Cloudflare's beta features and limits.
Manual approval workflow adds safety gate Chaos Mesh and Litmus don't prioritize.
Local-first Kubernetes workspace with native FluxCD and ArgoCD support.