Skip to content

Lens MCP Server#

Lens premium subscription feature. See details on Lens Pricing.

Lens MCP Server connects your AI assistants to Lens Desktop, enabling you to investigate and troubleshoot Kubernetes clusters using natural language. It implements the Model Context Protocol (MCP), an open standard by Anthropic donated to the Agentic AI Foundation (AAIF).

With the Lens MCP Server, you describe what you need in natural language, and the server runs read-only kubectl and PromQL commands on your behalf. You can troubleshoot cluster issues, investigate resource states, query Prometheus metrics, and correlate data across these sources to pinpoint problems that span multiple clusters or resource types.

Key capabilities#

  • List and manage cluster connections from AWS EKS, Azure AKS, local kubeconfig, and Lens Teamwork Spaces
  • Run read-only kubectl commands (get, describe, logs, top, and more) across connected clusters
  • Query Prometheus: list metrics, view alerts, get targets, execute range queries
  • Works with any MCP-compatible client: Claude Desktop, Claude Code, Cursor, Windsurf, ChatGPT, and others

  • Getting started

    Enable the MCP Server and connect your first AI assistant in under 5 minutes.

  • Kubernetes tools

    List clusters, connect to them, and run read-only kubectl commands.

  • Prometheus tools

    Query metrics, view alerts, inspect targets, and execute PromQL queries.

Requirements#

  • You have a Plus, Pro, or Enterprise subscription plan. For details, see Lens Pricing
  • A Lens Desktop version that includes the MCP Server. If the Lens MCP Server toggle is missing from Preferences > Integrations, update Lens Desktop to the latest release.
  • The Lens CLI is bundled with Lens Desktop and used automatically by the MCP Server. No separate installation is required.
  • An MCP-compatible AI assistant (Claude Desktop, Claude Code, Cursor, Windsurf, or other MCP client)

Limitations#

  • Read-only mode only. No write operations through the MCP server.
  • Google GKE integration is not yet available
  • Lens Desktop must be running for the MCP server to work