Radar

How It Works

1. 1

   Install

   A single ~30MB Apache 2.0 binary installable via curl | sh, Homebrew, Krew, Scoop, .rpm, direct download, a desktop app, or Helm in-cluster. Example: `curl -fsSL https://get.radarhq.io | sh && kubectl radar`.

2. 2

   Open your cluster

   Local mode opens like k9s and reads your kubeconfig; in-cluster mode runs as a Helm release and serves the UI through ingress for the whole team. Same features and UI in both modes.

3. 3

   See what's running

   Topology, events, Helm releases, GitOps state, image filesystems, and 44 audit checks — live, in one view. You read it in the UI; your agents read the same data over MCP.

Use Cases

• Troubleshoot Kubernetes incidents without kubectl: search across every cluster, jump to logs, rewind the event timeline, and share the link
• Manage a fleet of clusters from one search bar and compare workload versions side by side without kubeconfig hunting
• Get GitOps visibility with native ArgoCD and Flux sync state alongside the resources produced
• Audit Helm releases and roll back or diff between revisions
• Keep history of events and changes after Kubernetes deletes them (up to 30 days on Cloud)
• Let AI agents diagnose cluster issues via the built-in MCP server or Cloud AI investigations

Pricing

Radar OSS is free forever under the Apache 2.0 license — run it locally or in-cluster with no account or cluster limits. Radar Cloud is priced per cluster, never per seat: Cloud Free ($0, up to three hosted clusters), Team ($149 per cluster/month, starting price), Enterprise ($299 per cluster/month), and Custom (annual contract, contact sales). All paid plans include a 14-day free trial, annual co

Integrations

• Helm
• ArgoCD
• Flux
• Slack (alert delivery on Cloud)
• Webhooks (alert delivery on Cloud)
• MCP (Model Context Protocol) for AI agents
• SIEM export (Enterprise)
• Google and GitHub login (Cloud)