Skip to content
MCP server

KubeDrive

Let AI agents read your Kubernetes clusters through the KubeDrive desktop app

Runs locallyStatus and history
Install
Pricing
Freemium
Platforms
macOS, Windows, Linux
Runs
Locally
Version
0.6.6
Listing updated
3 days ago

About KubeDrive

KubeDrive is a freemium MCP server for developer tools. It runs locally on your machine and is started by your AI client.

The server is a local endpoint inside the KubeDrive desktop app, listening on 127.0.0.1 port 47011. Agents authenticate with a client token that starts with kd_, and they never receive a kubeconfig or a cluster token. KubeDrive answers using the connections it already holds: your kubeconfig, SSO sessions and RBAC permissions.

You choose which clusters are shared, and nothing is shared until you tick it. Each cluster has its own rules: limit reads to listed namespaces, hide Secrets entirely, and switch individual write actions on or off. Delete, exec and drain stay off until you enable them.

Reading is free for one agent at a time. Changes made by agents belong to the Pro plan, and every change waits for your Allow in a window that shows the diff. On a protected cluster you type the object's name to confirm. The setup guide covers Claude Code, Codex, Claude Desktop and Cursor.

Install

Run in a terminal

claude mcp add kubedrive --env KUBEDRIVE_MCP_TOKEN='<KUBEDRIVE_MCP_TOKEN>' -- npx -y https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb

Add to claude_desktop_config.json

{
  "mcpServers": {
    "kubedrive": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb"
      ],
      "env": {
        "KUBEDRIVE_MCP_TOKEN": "<KUBEDRIVE_MCP_TOKEN>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "kubedrive": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb"
      ],
      "env": {
        "KUBEDRIVE_MCP_TOKEN": "<KUBEDRIVE_MCP_TOKEN>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "kubedrive": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb"
      ],
      "env": {
        "KUBEDRIVE_MCP_TOKEN": "<KUBEDRIVE_MCP_TOKEN>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "kubedrive": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb"
      ],
      "env": {
        "KUBEDRIVE_MCP_TOKEN": "<KUBEDRIVE_MCP_TOKEN>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "kubedrive": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb"
      ],
      "env": {
        "KUBEDRIVE_MCP_TOKEN": "<KUBEDRIVE_MCP_TOKEN>"
      }
    }
  }
}

Configuration

  • KUBEDRIVE_MCP_TOKENThe client token from KubeDrive → AI Agents, after the server is turned on; starts with kd_Required
  • KUBEDRIVE_MCP_URLKubeDrive's MCP endpoint, only if its port was changed

Tools this server provides (3)

Taken from the project documentation; the running server may differ.

cluster_problems
Diagnoses problems across a whole cluster; refused when reads are limited to listed namespaces.
upgrade_check
Lists what the next Kubernetes version breaks.
rightsizing
Suggests resource right-sizing for workloads.

What it does

Local-only endpoint
The server runs on your computer at 127.0.0.1 and is switched on from the AI Agents window.
No kubeconfig for agents
Agents ask KubeDrive, which answers with its own connections, so no cluster token leaves the app.
Per-cluster rules
Restrict reads to listed namespaces, hide Secrets and choose which write actions are allowed.
Approval for changes
Each agent change waits for your Allow, with a typed object name on protected clusters.
Propose-only mode
An agent's change can be saved as a proposal instead of being applied.
Call indicator
A robot icon in the title bar shows server state and lights up at each agent call.

Good for

  • Ask why a deployment in staging is not ready
  • Find the pods that restarted most in the last hour
  • Check what a Kubernetes upgrade would break
  • Review rightsizing suggestions for a namespace
  • Give Claude Code or Codex cluster access without sharing a kubeconfig

Pricing

Freemium, from $8 a month

Reading free; Pro $8/mo or $79/yr adds agent changes; Team from $15/seat/mo

Prices change. Check kubedrive.com before you buy.

KubeDrive FAQ

Does the agent get my kubeconfig?

No. The agent never receives a kubeconfig or a cluster token; KubeDrive answers with its own connections.

Is it free?

Reading is free for one agent at a time. A second client and changes by agents need Pro, listed at $8 a month or $79 a year, with a 14-day trial.

Which clients work?

Claude Code, Codex, Cursor, Claude Desktop and other MCP clients; Claude Desktop connects through a small bridge called kubedrive-mcp.

Can an agent delete or drain without asking?

Delete, exec and drain are off on every cluster until you turn them on, and every change still waits for your Allow unless you set a rule for scale and restart.

Is KubeDrive free?

KubeDrive has a free plan, with paid plans for more usage or features. Paid plans start at $8 a month.

How do I add KubeDrive to Claude Code?

Run this command in your terminal: claude mcp add kubedrive --env KUBEDRIVE_MCP_TOKEN='<KUBEDRIVE_MCP_TOKEN>' -- npx -y https://github.com/aleksskoz/kubedrive/releases/download/v0.6.6/kubedrive-0.6.6.mcpb

What are the best alternatives to KubeDrive?

The closest alternatives to KubeDrive are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to KubeDrive

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools