DISCOVER THE FUTURE OF AI AGENTS

gcloud-mcp

Added May 8, 2026
Agent & Tooling
Open Source
TypeScriptNode.jsModel Context ProtocolAI AgentsCLIAgent & ToolingAutomation, Workflow & RPAProtocol, API & Integration

An official Google-maintained MCP server collection that allows AI assistants to manage Google Cloud resources, query observability data, and operate storage buckets via natural language.

gcloud-mcp is an open-source Model Context Protocol (MCP) server collection maintained by the official Google googleapis organization, designed to eliminate the barriers of complex CLI syntax in traditional cloud management. Built with TypeScript and running on Node.js (≥20), the project interfaces with mainstream AI clients (such as Gemini CLI, Claude Desktop, Cursor, and VS Code) via the stdio protocol.

Architecture

The project adopts a Monorepo structure managed by npm workspaces + Turborepo, divided into four independent server modules:

  • gcloud-mcp (Core): Wraps gcloud CLI execution with the run_gcloud_command tool, featuring built-in security filtering that blocks high-risk commands unsuitable for AI agents (e.g., commands accepting arbitrary input or launching interactive sessions).
  • observability-mcp: Provides tools for Cloud Logging, Cloud Monitoring, and Cloud Trace, covering list_log_entries, list_metric_descriptors, list_time_series, list_alert_policies, list_traces, get_trace, and more.
  • storage-mcp: Covers full-stack Cloud Storage management, including object-level operations (list/read/write/delete/copy/move/upload/download) and Bucket-level management (list/create/delete/metadata/labels/IAM), plus Storage Insights schema queries and configuration.
  • backupdr-mcp: Focuses on full CRUD operations for Backup and Disaster Recovery resources, including backup vaults, backup plans, data sources, and backup restoration.

This modular design allows users to mount and combine cloud management capabilities on demand.

Security Model

Permissions are strictly bounded by the locally authenticated gcloud account, with no separate auth configuration needed. Natively supports Service Account Impersonation for enterprise-level least-privilege control. The core server includes a denied commands blacklist to prevent AI agents from executing out-of-bounds operations.

Quick Start

# Gemini CLI / Gemini Code Assist (recommended)
npx @google-cloud/gcloud-mcp init --agent=gemini-cli

# Verify connection
gemini mcp list

For Claude Desktop / Cline / Cursor / VS Code, add the following to the respective MCP configuration file:

"gcloud": {
  "command": "npx",
  "args": ["-y", "@google-cloud/gcloud-mcp"]
}

Important Notice

The project is currently in Preview stage and may contain breaking changes. The README explicitly states that this project "provides a solution, not an officially supported Google product" and is not covered by Google Cloud Terms of Service.

Related Projects

View All

STAY UPDATED

Get the latest AI tools and trends delivered straight to your inbox. No spam, just intelligence.