AI-native control plane for Kubernetes

Operate Kubernetes
in plain English.

Klariq turns your cluster into a conversation. Ask, inspect, troubleshoot, generate YAML, and ship — without leaving the chat.

100% local LLM·No telemetry·Read-only by default·MIT licensed
klariq · cluster: prod-aks-01
Why is the payment service crashing?
Ask Klariq anything about your cluster…⌘K
30+
Resource views
12
AI-aware actions
<2s
Time-to-insight
0 ✨
Lines of YAML
How it works

From question to fix in three steps.

01

Ask anything

Type a question in plain English. “Why is checkout failing?” · “Show me cost by namespace.” · “Generate an Ingress with TLS.”

02

Klariq investigates

A local LLM picks the right MCP tools — kubectl, Prometheus, AlertManager — and assembles context across signals.

03

Get an answer + action

Diagnostics with citations, ready-to-apply YAML patches, cost numbers, security findings — all linked back to live cluster state.

What it does

Everything you need, none of the YAML rage.

AI Chat for Ops

Ask in plain English — Klariq inspects pods, deployments, services, and more in real time.

Live Observability

Cluster health, metrics, events, and logs unified into a single timeline view.

Security Posture

Score every namespace against CIS benchmarks, with one-click remediation hints.

Cost Insight

See spend per namespace and workload, with concrete savings recommendations.

YAML & Templates

Generate manifests from natural language or pick from a curated template gallery.

GitOps Aware

ArgoCD apps, Helm releases, and audit trail — all surfaced in one workspace.

Demo mode · no signup

See Klariq running on a sample cluster.

A pre-loaded production-style AKS workload with pods, deployments, alerts, costs, and ArgoCD apps — explore the full UX in your browser.

Open Demo
Try saying

What would you ask your cluster?

List all pods in CrashLoopBackOffWhy is the api-gateway slow?How much does staging cost this month?Generate a Postgres deployment with PVCShow security risks in productionScale frontend-web to 5 replicasWhich workloads are over-provisioned?Find pods running as rootAudit who deleted the order-service podDraft an HPA for analytics-worker
Built with
Next.js 14React 18TypeScriptTailwindframer-motionFastAPILangChainOllamaQwen 2.5MCPKubernetesPrometheusAlertManagerArgoCDHelm
About

I built Klariq because I've lived the pain.

Hi, I'm Mehmood — a Certified Kubernetes Administrator (CKA) and DevOps engineer who has spent more late nights than I'd like to admit staring at kubectl describe output, jumping between 6 terminals, and copy-pasting pod names into Grafana just to answer one simple question: “why is this thing broken?”

Every incident felt the same — context scattered across tools, runbooks out of date, YAML errors only revealing themselves at 2 AM. I kept thinking: there has to be a better way.

Klariq is my answer. An AI co-pilot that knows your cluster, speaks your language, and turns hours of debugging into a single conversation. Built by someone who's been on-call, for everyone who still is.

CKA CertifiedOn-call survivorYAML whispererCluster archaeologist2 AM responder
MU
Mehmood ul Hassan
CKA · DevOps Engineer · Cloud & Kubernetes

Building developer-friendly platforms on Kubernetes & Azure.

Contact

Let's talk Kubernetes, AI, or platforms.

Feedback, ideas, or just curious about the build? I'd love to hear from you.