All Enterprise features free for 3 months
Every signed-in user gets full Enterprise access to Kube Hammer until Oct 15, 2026. No credit card required.
Kube Hammer
Operations Engine
Pods, logs, shell, YAML and RBAC unified in one UI. Every change is snapshotted before it is applied. AI diagnosis uses the full cluster context for root-cause analysis. Runs offline and sends zero telemetry.
169
Marketplace & Plugins
Install plugins, templates and actions in one click — or write and publish your own. You decide how Kube Hammer grows.
Cluster Performance
kube-hammer
Command Palette
kube-hammer
eBPF Injection
kube-hammer
AI Assistant
kube-hammer
46+ official plugins, templates and actions. Filter by category, pricing and popularity; install and update with one click.
Develop, package and publish your own plugin to the marketplace with the Developer tab and SDK 2.0 — AST-scanned for safety.
Extend via Plugins, Templates and Actions. Pick exactly what you need across Free, Pro and Enterprise tiers.
Operational Capabilities
Find workloads, inspect state, open a shell, tail logs and edit manifests — without switching between tools.
See at a glance how your Ingress, Service, Deployment and Pods connect. Not just visual sugar — an interactive architecture graph designed for operational clarity.
Full xterm.js terminal for multi-container pods. Stream logs in real time with regex filtering — no extra window needed.
Two options: FizyonAI Cloud with end-to-end encrypted root-cause analysis, or bring-your-own-key Ollama for environments where no data may leave the network.
Every change made through Kube Hammer is snapshotted before it is applied. Deleted resources can be restored with one click. Timed operations auto-revert on timeout. The Timeline tab overlays your actions against live Kubernetes events so root causes are easy to find.
Manage nodes right from the UI: cordon / uncordon for maintenance, drain to evacuate workloads safely, top for live CPU/memory, list a node's pods, and a one-click Net Tools shell. When you need it, Capture Traffic (PCAP) spins up an ephemeral netshoot container to capture live network traffic from any Pod or Node (full packet payloads, up to 60 seconds) and downloads a .pcap ready for Wireshark.
Small but powerful utilities that run quietly in the background while you work — just like traffic capture.
Export cluster events as a one-time snapshot or watch and stream them continuously. JSON, CSV or plain text — with local-file output and size-based rotation.
Keeps your last copied items (up to 100) so you can re-copy a command, certificate or log output with one click, and reorder them however you like.
Scheduled etcd snapshots with retention, compression, encryption and multi-cloud storage — protecting your cluster state for disaster recovery.
Tailor the row actions for every resource type. Reorder them with arrow buttons, pick how many stay visible as primary with a slider — the rest fall into the "⋮" overflow menu. Pin the ones you use most as always visible. Both built-in actions and marketplace plugin actions are listed here, and your preference is saved per resource type.
Define exactly which kubectl commands run for each resource type. Add your own actions using {name} and {namespace} placeholders, and set each one's name, description, icon, confirmation prompt, quick action and mutating (list-refresh) behaviour. Choose where the output goes: dialog, popup, bottom/right panel, detached window or external terminal.
Decide exactly which fields appear in your resource tables. Build custom column templates per resource kind, turning any field of the raw Kubernetes object into a column via data paths like _raw.status.podIP. Show/hide columns, reorder them with arrow buttons, and align them left/center/right. Shape values with expression functions such as prettySize, age, percentage, colorize and conditional. Start from built-in views (Compact, Detailed, per-kind), set one as default, clone them, and import/export as JSON.
Almost every capability in Kube Hammer is a plugin you can toggle on or off. Enable 50+ built-ins from the Plugin Manager: Flux GitOps (whose Drift Defender flags when the cluster diverges from Git and re-syncs with flux reconcile), Helm, the Trivy CVE scanner, TLS Detective, eBPF Injection, OpenTelemetry, etcd Backup, Chaos Simulator and more. Plugins aren't just tools — they add sidebar views, table columns (like Root Controller and Controlled By, derived from ownerReferences), row actions and YAML spec editors. Install from the marketplace, side-load your own from disk, or write one in minutes with the SDK.
Describe what you need in plain language and the Recipe Builder reads the core engine's behaviour plus every installed plugin capability to generate a ready-to-run recipe or capability plan that combines them into a single operational workflow. It uses your configured AI endpoint (FizyonAI Cloud, Ollama or OpenAI-compatible), injecting all plugin capabilities into the LLM context. Start from pre-written templates — Safe rollout, Troubleshoot, Probe snooze, Change flow — so you extend Kube Hammer not only with plugins, but with natural language.
Built-in AI
Kube Hammer's AI assistant reads your logs and events, finds the root cause, and either suggests kubectl commands or safely runs them for you — all with the context of the resources you've selected, right inside the panel.
Interprets raw logs and events to surface the root cause — CrashLoopBackOff, OOMKilled, a failing probe, a missing secret — then walks you through step-by-step remediation.
In Agent mode the assistant auto-runs whitelisted kubectl commands and iterates on the output. In Ask mode it only proposes commands — nothing runs until you press Run.
With Auto-send on, every executed command's result is fed back to the assistant so it advances the diagnosis on its own (up to 3 follow-up steps). With it off, you forward output manually.
One click on a selected resource for Explain, Debug, Check logs, Events, Generate YAML and Describe. Each button sends a ready-made diagnostic prompt to the assistant.
Choose Minimal, Standard or Deep. Minimal sends only core context; Standard adds open panels and filters; Deep also attaches recent errors, events, command outputs and selected log snippets.
Namespace-aware by design. Pods, deployments and other resources you select in the table — multiple kinds at once — enter the assistant's context, and you can see exactly what's sent.
The kubectl commands the assistant generates arrive inline with Run and copy buttons. They execute safely in-app, gated by an allowed-command list and a 15-second timeout.
Open independent, named chat sessions (up to 30); past messages (up to 100) are stored on your device. Keep context tidy with New Chat, Sessions and Chat History.
Answers stream live as markdown. Expand a separate 'View thinking' section to inspect the model's reasoning steps, or just read the final result.
Your choice of provider: FizyonAI Cloud · local Ollama · OpenAI-compatible endpoints (BYOK). Before any request leaves your machine, secrets, base64 values and Kubernetes secret fields are automatically redacted.
Full Capabilities
A growing library of 45+ built-in and marketplace capabilities spanning operations, AI, security, FinOps, GitOps, and observability.
Comparison
From pod management to embedded AI, from snapshot-based safe operations to air-gapped compliance — see what Kube Hammer solves on its own for day-to-day operations.
| Capability | Kube Hammer | kubectl | Lens | k9s |
|---|---|---|---|---|
| Visual UI + integrated terminal & logs | ||||
| Relational resource graph (DAG) | ||||
| Built-in AI diagnostics (Cloud / local Ollama) | ||||
| Snapshot + auto-revert safe operations | ||||
| Offline / air-gapped operation | ||||
| Zero telemetry (off by default) | ||||
| Plugin SDK + marketplace |
Comparison is based on publicly available feature information and is for reference only. Every tool is strong in its own domain.
Snapshots before every change, conflict checks on revert, AST-scanned plugin sandboxing. Your credentials never leave your machine. Built to the standards DevSecOps teams demand.
Every marketplace plugin passes Abstract Syntax Tree (AST) static analysis. Obfuscated payloads and hidden exfiltration attempts are automatically blocked.
Plugins run in strict isolation. If a UI-only plugin tries to access the network, the Kube Hammer core blocks it at runtime.
MitM-resistant. Enterprise tool licenses use RSA-signed JWTs verified against hardcoded public keys.
Zero inbound ports required. Supports enterprise SSO and offline install for heavily audited environments. Never sends telemetry by default.
ClusterDiffEngine structurally compares every resource between two cluster contexts (e.g. prod vs staging). ResourceSnapshotStore keeps a 24-hour ring buffer of state, answering "what did this pod look like 30 minutes ago?" exactly.
No DSL to learn. The plugin SDK is built on React and Zustand. It ships a Worker pool for heavy work, a ServiceBus for cross-plugin messaging and an ActionPipeline for intercepting commands.
Inject React components into well-defined points of the UI.
Sandboxed architecture. Declare network access explicitly in the manifest.
Add your custom resource types to the left nav as if they were native.
Decoupled, safe communication between plugins.
Community is free forever for individuals. Choose Enterprise for strict governance, air-gapped architecture and centralized management.
Core features for individuals and small teams.
Advanced analytics, cross-device sync and architecture diagnostics.
For heavily audited organizations and air-gapped architectures.
~/.kube/config and talks directly to your cluster API server. No credentials are sent to Fizyonops or any third-party service. Telemetry is off by default.Native installers for Linux, macOS and Windows. Drop in your kubeconfig and manage clusters locally with scoped tokens. No signup wall — just pick your platform and download the latest signed build.
Signed installers for Linux (.deb / .rpm), macOS (Apple Silicon and Intel) and Windows (.exe); Android on Google Play.
Bring teammates onboard with role-based invites and manage cluster access from a single portal.
Pick your platform and start in minutes. No signup, no email — just the latest signed build.
By downloading, you agree to our Terms and Privacy Notice.