KubeDrive 0.6 — Upgrade check, rightsizing and a map of your namespace

Kubernetes,
at native speed.

KubeDrive is a fast, native client for your clusters on macOS, Windows and Linux. See what needs attention, dig into any resource, follow logs and open a shell — without leaving the keyboard.

Download for macOS

brew install --cask aleksskoz/tap/kubedrive

macOS 13 or later · Apple silicon and Intel · signed and notarized by Apple

Download for Windowsx64 · Arm64

Windows 10 1809 or later, Windows 11 · not code-signed yet: More info → Run anyway

Download .deb Download .rpmArm64: .deb · .rpm

sudo apt install ./kubedrive_0.6.7_amd64.deb

Debian 13, Ubuntu 24.04, Fedora 40 and newer · GTK 4 and WebKitGTK 6.0 come as dependencies

Free forever for two clusters · Pro free for 14 days, no card

KubeDrive diagnosing a crash-looping pod by rules: the cause, the last lines of its log and a button to open the YAML

Native app for macOS, Windows and Linux
Your kubeconfig, your credentials
EKS · GKE · AKS · DOKS · OpenShift · kind
Cluster data never reaches our servers

Demo

KubeDrive in two minutes.

A crash loop found and explained, logs merged across a Deployment, port forwards, the upgrade check, and Claude Code reading the cluster through KubeDrive. Recorded on a local kind cluster, no cuts.

Watch on YouTube · More videos

Features

Everything you reach for, a keystroke away.

AI assistant

Why a pod is failing, found by rules first, without AI: the cause, the evidence from events and logs, and a fix. Ask the assistant for more, with your own key: Claude, OpenAI or a local model.

Upgrade check

What breaks on the next Kubernetes version, before you upgrade: removed APIs in Helm releases and applied manifests, clients still calling them, nodes too old.

Rightsizing

Requests and limits from what containers really use, by the cluster's Prometheus: what is at risk of being killed or throttled, and what reserves room it never uses.

Map

How a namespace's Ingresses, Services, workloads, pods and config connect, and which link leads nowhere: a Service with no pods, a missing Secret or ConfigMap.

Prometheus dashboards

CPU, memory and network for every pod and node from the cluster's own Prometheus, PromQL, your dashboards and the alerts that fire.

Helm charts

Browse chart repositories and OCI registries, install with a dry run first, upgrade with a manifest diff, roll back in a click.

AI agents

Claude Code, Cursor and Codex read the clusters you choose through KubeDrive's local MCP server; every change waits for your Allow.

Live resource tables

Every resource type, streamed live, grouped by namespace, with logs, a shell and YAML beside the list.

Keyboard first

⌘K for everything: switch clusters, namespaces and resources without the mouse.

Production safety

Safe on production, by default.

The expensive mistakes happen on production. KubeDrive guards against them before you set anything up.

Protected without setup
A context named prod, prd, production or live is protected from the first launch. Set it per cluster: auto, always or never.
Type the name to change it
On a protected cluster every change asks you to type the object's name. What your RBAC role does not allow is shown, but disabled.
Always clear where you are
A red frame around the window, a red dot on the cluster's tab and a PROD label in the cluster switcher.
AI agents only where you allow
Agents see only the clusters you share, none by default. Changes stay off until you turn them on, and each one waits for your Allow.

How protection works

All your clusters, with their health, on one screen.

KubeDrive reads your kubeconfig and checks every context: reachable, needs a VPN, needs an SSO login. Open one and you land on what matters.

Connect your first cluster

The clusters overview: each context with its status, and the problems inside the cluster that is open

See what each pod uses, from the Prometheus you already run.

KubeDrive finds the cluster's Prometheus and charts CPU, memory and network by namespace, pod and node. Write PromQL, keep your own dashboards and see the alerts that fire.

Metrics and PromQL

A namespace dashboard with CPU, memory and network charts for each pod

Helm charts, installed with a dry run first.

Add chart repositories and OCI registries, read a chart's README and default values, install with your values, upgrade with a manifest diff and roll back when a release goes wrong.

Helm in KubeDrive

The Helm chart catalog with the cert-manager chart's README open

Let your AI agent read the cluster, safely.

Turn on KubeDrive's local MCP server and Claude Code, Cursor or Codex read the clusters you pick, with your credentials. Secrets are never shared, and every change an agent proposes waits for your Allow.

Connect an agent

The AI Agents window: the MCP server running, one cluster shared and changes off until turned on

Pricing

Free for daily work. Pro when you scale.

MonthlyYearly save 18%

Free

$0forever

For one developer and a couple of clusters.

Download

2 clusters open at a time
3 active port-forwards
Helm releases, read-only
Basic CPU and memory charts
AI “Why?” and log summaries with your own key
Diagnosis, upgrade check, rightsizing and map

Pro

Team

For teams that share clusters. From 5 seats, $15 a seat a month, billed yearly.

Buy Team

Pick the number of seats at checkout. The owner hands them out in the app.

Everything in Pro, on every seat
Seats given and taken in the License window
One yearly payment for the whole team
Priority support

Prices in US dollars. VAT included in the EU and UK; sales tax added where applicable. Polar, our reseller, may show them in your currency at checkout. Students and open-source maintainers get Pro free — write to us.

Open your clusters in KubeDrive.

It reads the kubeconfig you already have. Nothing to set up on the cluster.