Cloud Run MCP Server

Official

MCP server to deploy apps to Cloud Run

How to install ↓ Compare GitHub source repository ↗

Category: Cloud Storage #13 of 15
Stars: ★ 633 Very popular
Transport: stdio (local process)
Runtime: Node.js · Docker
Credentials: API key / credential required
License: Apache-2.0
Last commit: 3 days ago
Tools: 7

47FMRS · D

A powerful MCP server for Cloud Run deployment, offering a variety of tools and prompts. It is great for AI-assisted workflows but requires careful configuration of authentication and security settings.

Strongest · Documentation 13/20 Weakest · Security and permissions 3/20

Reliability

6/20

Security and permissions

3/20

Maintenance

13/20

Documentation

13/20

Setup experience

12/20

Why each score

Reliability 6/20: Only the README and repository metadata were available; the server manifest was not cached and no source, CI workflows, or committed tests could be inspected. The README lists seven tools and two prompts with npx/docker/node startup options, but actual tool behavior, error handling, and dependency controllability cannot be confirmed. The local-only availability of deploy-local-folder/list-projects/create-project adds conditional behavior. The static calibration cap applies because there is no verifiable execution evidence, so reliability is 6.
Security and permissions 3/20: Credentials are handled via gcloud ADC / GOOGLE_APPLICATION_CREDENTIALS, the remote mode emphasizes IAM authentication and warns against unauthenticated use, and OAuth is documented. However, SKIP_IAM_CHECK defaults to true, which makes Cloud Run services public; ENABLE_HOST_VALIDATION is disabled by default, leaving DNS rebinding protection off; and the deploy-type tools trigger deployment side effects without any confirmation mechanism described in the README. This touches the red line of unconfirmed deploy side effects, so security is capped in the 0-4 range at 3.
Maintenance 13/20: The repository is not archived, is Apache-2.0 licensed, and has clear GoogleCloudPlatform ownership; the README is actively maintained with OAuth, skills, and Gemini CLI additions. However, there is no evidence of release cadence, commit frequency, dependency updates, or a security-response channel, and the 22 open issues lack visible handling details. Maintenance is active but governance/versioning evidence is incomplete, so the score is 13.
Documentation 13/20: The README covers local, remote, Docker, and OAuth installation paths, client configuration examples, an environment variable table, a tool list, and prompt descriptions, with links to client-specific configuration docs. It lacks per-tool parameter schemas/input fields, quotas, cost implications, troubleshooting guidance, and reviewable example outputs. The documentation is usable but has hidden assumptions and troubleshooting gaps, so it scores 13.
Setup experience 12/20: The local setup path is clear: install gcloud, log in, and add the MCP JSON; Docker and remote options are also documented, with configuration examples for several mainstream clients. However, setup depends on gcloud authentication, application default credentials, and optional OAuth/proxy steps, making it manual and somewhat fragile. Without CI/test evidence that these configurations reproducibly connect, the static calibration cap applies; setup is 12.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesUses the networkChanges third-party account data

Best for

Developers and teams wanting to integrate deployment into AI workflows
Users already using Google Cloud and Cloud Run
AI agents that need to quickly prototype or deploy applications

Not for

Deployments requiring local hosting or other cloud provider support
Users wanting fine-grained IAM control without skipping permission checks
Seeking GitOps or fully automated CI/CD without MCP/CLI

Required permissions

Requires Google Cloud IAM credentials to access Cloud Run APIs.
Authentication via `gcloud auth login` and `gcloud auth application-default login`.
Deployment requires appropriate Cloud Run permissions like `run.services.create` and `run.services.update`.
Project creation requires permissions like `resourcemanager.projects.create`.

Risks and side effects

Using remote MCP server without authentication may lead to unauthorized access.
IAM checks are skipped by default (`SKIP_IAM_CHECK=true`), potentially allowing overly permissive access.
Deploying file contents may expose sensitive data if provided in prompts.
DNS rebinding risk unless host validation is enabled (`ENABLE_HOST_VALIDATION`).

Setup

Before you start

Runtime：Node.js · Docker

GOOGLE_APPLICATION_CREDENTIALS optionalsecret Docker setup only: path to the application default credentials JSON produced by gcloud auth application-default login.

Other optional settings (6)

GOOGLE_CLOUD_PROJECT optional Default GCP project ID for deployments; find it in the Cloud Console or via gcloud config; optional.

GOOGLE_CLOUD_REGION optional Default Cloud Run region, e.g. us-central1; optional.

DEFAULT_SERVICE_NAME optional Default service name when none is provided; optional.

SKIP_IAM_CHECK optional Skips IAM permission checks; true by default, set to false to enable checks; optional.

ENABLE_HOST_VALIDATION optional Enables Host header validation against DNS rebinding; disabled by default; optional.

ALLOWED_HOSTS optional Comma-separated allowed Host headers when validation is enabled; defaults to localhost,127.0.0.1,::1; optional.

Install Google Cloud SDK and authenticate: gcloud auth login and gcloud auth application-default login.
Configure MCP client with npx command: npx -y @google-cloud/cloud-run-mcp.
Optionally set environment variables like GOOGLE_CLOUD_PROJECT, GOOGLE_CLOUD_REGION, and DEFAULT_SERVICE_NAME.

claude_desktop_config.json

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

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

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

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add cloud-run -- npx -y @google-cloud/cloud-run-mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After connecting, the client's tool list should show deploy-file-contents, list-services, and the other tools; then prompt like “list the Cloud Run services in this project” — a returned service list confirms authentication and connection work.

Troubleshooting

Ensure gcloud is installed and authenticated properly.
Verify default project and region settings.
Check `SKIP_IAM_CHECK` setting if encountering permission errors.
For remote setup, ensure the Cloud Run MCP server is deployed in the same project.
If connection issues, run `gcloud auth application-default login`.

Things to try

Once connected, you can ask your AI assistant things like:

List all Cloud Run services in project my-project in region us-central1
Deploy the current directory as a Cloud Run service named hello-world
Show me the logs and error messages for the hello-world service
Get the details of the hello-world Cloud Run service

Tools 7

deploy-file-contents writes

Deploys files to Cloud Run by providing their contents directly.

list-services read-only

Lists Cloud Run services in a given project and region.

get-service read-only

Gets details for a specific Cloud Run service.

get-service-log read-only

Gets Logs and Error Messages for a specific Cloud Run service.

deploy-local-folder writes

Deploys a local folder to a Google Cloud Run service.

list-projects read-only

Lists available GCP projects.

create-project writes

Creates a new GCP project and attach it to the first available billing account.

Use cases

Deploy applications from AI-assisted IDEs (e.g., Cursor) or desktop AI apps (e.g., Claude)

AI-agent-driven deployment via Gemini CLI or agent SDKs

Automate Cloud Run service management in cloud environments

Supported clients

Gemini CLI

Claude DesktopPartial support

VS CodePartial support

CursorPartial support

WindsurfPartial support

AntigravityPartial support

Listed from the project's documentation, not tested by this site.

Overview

The Cloud Run MCP server enables MCP-compatible AI agents to deploy apps to Google Cloud Run. It provides tools to deploy files or local folders, list and manage services, and fetch logs. Additionally, it includes prompts for common tasks, and advanced tools for project management.

Similar servers

Google Search Console MCP Server 69 · C

Comprehensive access to Google Search Console via MCP, with advanced analytics and quick win detection.

★ 277 · Tools 1 Compare with this →

Nextcloud MCP Server 67 · C

A production-ready MCP server that connects AI assistants to your Nextcloud instance.

★ 381 · Tools 13 Compare with this →

reMarkable MCP Server 65 · C

Use your reMarkable tablet as a second brain for AI assistants. Read documents, browse files, extract text, and OCR handwriting.

★ 243 · Tools 13 Compare with this →

Microsoft 365 MCP Server 61 · C

An MCP server for interacting with Microsoft 365 and Office services through the Graph API.

★ 1.0k · Tools 10 Compare with this →

Compare with other MCP servers

Source revision 18588d9468cf Data synced 2026-10-12 Read the FMRS scoring method