Roadmap & vision

Where kprompt is headed: AI SRE investigation, trust loops, multi-cluster, and what stays out of scope. Honest shipped vs exploring.

Positioning

kprompt is an intent compiler today: natural language → reviewable PlanResult → approve → apply. The long-term bet is AI SRE — thinking about the cluster (investigate, why, timeline, blast radius, verify) while keeping the same approval contract. Not a free-form chat REPL, not a silent analyzer daemon, not a hosted fleet scanner. Read Building AI SRE in Public, Intent compiler, not chat, Kubernetes AI tools compared, and Beyond AI kubectl for public positioning.

Shipped now

  • Plan → safety → approve → apply (deploy, scale, rollback, named delete)
  • Deep explain, logs, describe, discovery-backed get/list
  • Helm, Argo Workflows, Prometheus, OTel, Grafana, multi-tool routes
  • Tekton, KEDA, Istio (read-first), Crossplane, Flux/Argo CD GitOps
  • Optimize-cluster report (idle / rightsizing / HPA) and service dependency graph
  • Context aliases, doctor, Homebrew, optional Team login / policy / audit
  • Local read-only inventory via kprompt dash (localhost)

Full integration detail lives on Integrations — not duplicated here.

Building next — trust & AI SRE

These deepen “think about the cluster” without leaving the plan/approve loop. Treat as in progress — not a public ship date.

  • Blast-radius preview on mutating plans (who/what is affected)
  • Post-apply verify — confirm the goal after --wait, not only “applied”
  • investigate — multi-hop RCA (ingress → Service → Endpoints → Pods → events → logs → mesh/DNS)
  • why — causal state trees (e.g. Pending → affinity → no GPU nodes)
  • timeline — incident chronology from events, rollouts, HPA (+ optional metrics)
  • audit / cleanup / drift — hygiene scans and GitOps drift reports with optional approved remediations
  • GitOps PR mode — open/update a PR instead of live mutate when Flux/Argo CD is in play

Exploring later

  • Proactive local watch (opt-in): signal → “latency up; investigate?” — never auto-mutate
  • Local remember / session digests (facts stay on your machine by default)
  • Multi-cluster: contexts inventory, read fan-out, explicit per-cluster mutate safety; org registry metadata without uploading kubeconfig
  • Team web polish on app.kprompt.ai (policy, audit, Insights) — CLI stays free; nothing to buy today
  • Workflow recipe packs (harden production, Ingress → Gateway API) as curated plan chains

What we will not do

  • Silent apply across clusters from one --approve
  • Upload kubeconfig / cluster credentials to the control plane
  • Replace Lens/Headlamp as a hosted live cluster browser
  • Compete on free-form agent chat with kubectl-ai
  • Claim K8sGPT analyzer parity as a fleet scanner

Shape priorities

Star issues and PRs on GitHub. The CLI stays Apache-2.0 and experimental — review every plan before apply. Start with Quickstart when you are ready to try it.