GCP MCP Server

Manage Google Cloud Platform infrastructure from your AI assistant. The GCP MCP server enables natural language management of Google Cloud resources, from Compute Engine instances to Cloud Run services, GKE clusters, and Cloud Storage.

This server is designed for cloud engineers managing multi-project GCP environments, developers deploying serverless applications on Cloud Run, and platform teams maintaining GKE clusters across their organization.

What it Does

• Compute Management: VM instances, instance groups, and autoscaling
• Container Services: GKE clusters, Cloud Run services, and Cloud Functions
• Storage: Cloud Storage buckets, Cloud SQL, and Firestore
• Networking: VPC, load balancers, and Cloud CDN
• Monitoring: Cloud Logging and Cloud Monitoring integration

Installation

# Using npx
npx -y @anthropic-ai/mcp-server-gcp

# Using Docker
docker run -i --rm -v ~/.config/gcloud:/root/.config/gcloud mcp/gcp

Configuration

{
  "mcpServers": {
    "gcp": {
      "command": "npx",
      "args": ["-y", "@anthropic-ai/mcp-server-gcp"],
      "env": {
        "GCP_PROJECT": "my-project-id",
        "GOOGLE_APPLICATION_CREDENTIALS": "${HOME}/.config/gcloud/application_default_credentials.json"
      }
    }
  }
}

Available Tools

Common Workflows

Investigating a Production Issue

1. Check which Cloud Run services are showing high CPU or errors
2. Query Cloud Logging for recent error logs
3. Inspect the affected service's configuration and scaling settings
4. Check GKE cluster status if the service runs on Kubernetes
5. Ask "Summarize what's failing and recommend fixes based on the logs"

Cost Optimization Review

1. List all Compute Engine instances across projects
2. Identify idle or underutilized instances
3. Check Cloud SQL instances for unnecessary backups or over-provisioning
4. Review Cloud Storage buckets for unused or oversized objects
5. Generate a cost optimization report with recommendations

Infrastructure Audit

1. List all resources across compute, storage, and networking
2. Check IAM permissions on critical resources
3. Identify publicly accessible storage buckets
4. Review VPC firewall rules for overly permissive configurations
5. Generate a security posture summary

Example Prompts

• "List all GKE clusters in my project"
• "Show me Cloud Run services with high CPU"
• "What Cloud Storage buckets exist?"
• "Check Cloud SQL instances for backups"
• "Show logs from the backend service"
• "Find all resources tagged with environment:production"

Troubleshooting

Prerequisites

• Google Cloud SDK (gcloud) installed
• Application Default Credentials configured
• Appropriate IAM permissions for target resources