jackwotherspoon/mcp-on-cloudrun
If you are the rightful owner of mcp-on-cloudrun and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
The Model Context Protocol (MCP) is a protocol designed to connect prompts, resources, and tools to a Language Model (LLM), facilitating enhanced interaction and functionality.
The Model Context Protocol (MCP) is a versatile protocol that enables seamless integration of prompts, resources, and tools with Language Models (LLMs). Originally designed to operate on a client-server architecture with local servers using the `stdio` transport, MCP has evolved to support remote server operations through additional transports like `sse` and `streamable-http`. This evolution has made MCP particularly popular in the realm of AI-powered Integrated Development Environments (IDEs) and agentic applications that require connectivity to external tools and resources. By deploying MCP servers remotely, users can leverage the scalability, centralized management, and security features offered by platforms like Google Cloud Run. This setup not only enhances the collaborative potential of MCP servers but also ensures secure and authenticated access, preventing unauthorized usage.
Features
- Scalability: MCP servers can scale automatically based on demand when deployed on platforms like Google Cloud Run.
- Centralized Management: A single MCP server can be shared among team members, ensuring consistent updates and functionality.
- Security: Enforced authentication ensures that only authorized users can access the MCP server, enhancing security.
- Remote Operation: Supports remote server operation through `sse` and `streamable-http` transports.
- Integration: Facilitates integration with various tools and resources for enhanced LLM interaction.
Usages
usage with Google Cloud Run
bash gcloud run deploy mcp-server --no-allow-unauthenticated --region=us-central1 --source .
usage with Cloud Build and Artifact Registry
bash gcloud builds submit --region=us-central1 --tag us-central1-docker.pkg.dev/$PROJECT_ID/remote-mcp-servers/mcp-server:latest
usage with Cloud Run proxy
bash gcloud run services proxy mcp-server --region=us-central1
usage with FastMCP client
bash uv run test_server.py
Tools
add
Takes two numbers and returns their sum.
subtract
Takes two numbers and returns their difference.