Google Cloud Run MCP server
Google Cloud's official Cloud Run MCP server: deploy local code or file contents to Cloud Run, list and inspect services, and read service logs from your agent.
Cloud Run is Google Cloud's serverless container platform that runs your code in fully managed containers, scaling automatically from zero. The official Cloud Run MCP server, published by Google Cloud, lets an agent deploy applications to Cloud Run and operate them directly — so a coding assistant can ship the project it just wrote and then read its logs without you switching to the gcloud CLI or the console.
The server exposes seven tools. deploy-local-folder deploys a folder from your machine and deploy-file-contents deploys code by passing file contents directly (the path used when the server runs remotely), each turning source into a running Cloud Run service. list-services enumerates the services in a project and region, get-service returns the configuration of a single service, and get-service-log fetches its logs and error messages so the agent can diagnose a bad deploy. list-projects and create-project (local-only) manage the surrounding GCP project, with create-project attaching a new project to your first available billing account. It runs over stdio with npx @google-cloud/cloud-run-mcp and authenticates through the Google Cloud SDK using your gcloud Application Default Credentials; GOOGLE_CLOUD_PROJECT and GOOGLE_CLOUD_REGION set defaults. The server can also be deployed remotely onto Cloud Run itself behind IAM authentication, and installs as a Gemini CLI extension.
Quick install
Copy-paste configs are provided for all 8 supported clients. Pick your client below.
Available tools
| Tool | Description |
|---|---|
| deploy-file-contents | Deploy code to a Cloud Run service by passing the file contents directly. Used when the server runs as a remote MCP server. |
| list-services | List the Cloud Run services in a given project and region. |
| get-service | Get the configuration details for a specific Cloud Run service. |
| get-service-log | Get logs and error messages for a specific Cloud Run service. |
| deploy-local-folder | Deploy a folder from the local filesystem to a Cloud Run service. Available when the server runs locally. |
| list-projects | List the available GCP projects. Available when the server runs locally. |
| create-project | Create a new GCP project and attach it to the first available billing account. Available when the server runs locally. |
Required configuration
- GOOGLE_CLOUD_PROJECTOptional
Default GCP project ID.
- GOOGLE_CLOUD_REGIONOptional
Default Cloud Run region (e.g. us-central1).
- DEFAULT_SERVICE_NAMEOptional
Default Cloud Run service name.
What you can do with it
Deploy the app you just built
Ask the agent to ship the current project to Cloud Run: deploy-local-folder packages and deploys the working directory, then get-service returns the live URL and configuration — no Dockerfile wrangling or gcloud commands required.
Diagnose a failing service
Use list-services to find the service, get-service to read its current revision and settings, and get-service-log to pull the error messages behind a 500 — letting the assistant point at the root cause and redeploy a fix.
FAQ
- Is it free?
- The MCP server is open source and free to run. It deploys to Cloud Run, which has a generous free tier and otherwise bills for the compute, requests, and networking your services use against your Google Cloud account.
- Does it support remote/OAuth?
- The default setup runs locally over stdio with npx @google-cloud/cloud-run-mcp and authenticates with your gcloud Application Default Credentials. You can also deploy the server itself onto Cloud Run as a remote MCP server protected by IAM authentication.