This repository provides resources for Google's Model Context Protocol (MCP), enabling developers to integrate various Google services and models into their applications. It offers lists of managed remote MCP servers and open-source implementations, along with deployment guidance and examples for Google Cloud, targeting developers building AI-powered agents and applications.

How It Works

The project centers around the Model Context Protocol (MCP), which facilitates communication with AI models and services. It distinguishes between Google-managed remote MCP servers (like Google Maps, BigQuery) and open-source MCP servers that can be self-hosted or deployed on Google Cloud platforms such as Cloud Run and GKE. This approach allows for flexible integration of diverse Google functionalities, from Workspace and Cloud Storage to generative models and security tools.

Quick Start & Requirements

• Deployment: Guidance and examples are provided for deploying MCP servers on Google Cloud Run and Google Kubernetes Engine (GKE).
• Examples: Includes a sample agent ("Launch My Bakery") built with the Agent Development Kit (ADK).
• Documentation: Links to official documentation, blog posts, and codelabs for setup and deployment are available within the repository.
• Prerequisites: Deployment typically requires a Google Cloud account and familiarity with services like Cloud Run or GKE. Specific MCP server examples might have additional dependencies.

Highlighted Details

• Comprehensive list of both Google-managed remote MCP servers and self-hostable open-source MCP servers.
• Extensive examples and codelabs demonstrating agent development (using ADK) and MCP server deployment on Google Cloud.
• Integration support for a wide array of Google services, including Workspace, Cloud Storage, Genmedia (Imagen, Veo), Security Command Center, and more.

Maintenance & Community

The repository welcomes contributions, including bug reports, feature requests, and code contributions, with specific guidelines provided. No specific community channels (like Discord/Slack) or roadmap links are explicitly mentioned in the README.

Licensing & Compatibility

Licensed under the Apache 2.0 License. Generally permissive for commercial use, but the project is intended for demonstration purposes and is not an officially supported Google product.

Limitations & Caveats

Explicitly stated as "not an officially supported Google product" and intended "for demonstration purposes only." Not eligible for the Google Open Source Software Vulnerability Rewards Program. No specific performance benchmarks or detailed architectural deep-dives are presented in the README.