Skip to content

Summary

Google Cloud's MCP server and Gemini CLI extension that lets AI agents deploy code to Cloud Run, list services and read their logs.

Features

  • Deploys a local folder or file contents to a Cloud Run service
  • Lists services and fetches service logs and errors
  • Can create a GCP project and attach billing when run locally
  • deploy and logs prompts for the current working directory
  • Also ships as a Gemini CLI extension

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it, filling in the environment variables with your own values.
  3. Restart the agent — the "Cloud Run MCP" tools will be available.
{
  "mcpServers": {
    "cloud-run-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@google-cloud/cloud-run-mcp"
      ],
      "env": {
        "GOOGLE_CLOUD_PROJECT": "<your-project-id>",
        "GOOGLE_CLOUD_REGION": "us-central1"
      }
    }
  }
}

Description

Cloud Run MCP is Google Cloud's server for deploying applications to Cloud Run from an AI agent. Ask Gemini CLI, an AI IDE or a desktop assistant to deploy the current folder, and it ships the code to a Cloud Run service; it can then list services, describe one and fetch its logs and errors. (It deploys apps to Cloud Run; hosting MCP servers on Cloud Run is covered separately in Google's documentation.)

Tools and prompts
  • deploy-file-contents deploys files passed directly as contents.
  • list-services, get-service and get-service-log inspect what is running in a project and region.
  • Run locally, it adds deploy-local-folder, list-projects and create-project, which creates a GCP project and attaches it to the first available billing account.
  • Prompts deploy and logs act on the current working directory by default.
Setup

Locally it runs with npx -y @google-cloud/cloud-run-mcp (or Docker) after gcloud auth login and gcloud auth application-default login. Defaults come from GOOGLE_CLOUD_PROJECT, GOOGLE_CLOUD_REGION and DEFAULT_SERVICE_NAME. It also installs as a Gemini CLI extension, and can itself be deployed to Cloud Run as a remote server, reached through a local Cloud Run proxy with IAM authentication (optional OAuth bearer-token checks are also available). By default it skips the IAM check so deployed services are public; set SKIP_IAM_CHECK=false to change that.

Who it is for

Developers on Google Cloud who want an agent to take code from the editor to a running serverless service. Apache-2.0.

Covered in the Weekly

MCP: Cloudflare

by Cloudflare

Cloudflare's catalogue of managed remote MCP servers — docs, Workers bindings and builds, observability, Radar, Browser Run, AI Gateway, CASB and more, each on its own OAuth endpoint.

Browse all MCP servers →