google/mcp is a collection of Model Context Protocol servers and deployment guidance from Google.
The repository addresses the need to connect large language models with Google Cloud services and data sources through the Model Context Protocol standard. It provides both remote MCP servers managed by Google that integrate with services like BigQuery, Cloud SQL, Firestore, and Compute Engine, as well as open-source MCP servers that can run locally or be deployed to Google Cloud. The remote servers are accessed via endpoints and cover a range of Google Cloud products in general availability and public preview, while the open-source servers offer flexibility for local deployment or custom cloud configurations.
Developers should choose this collection if they are building AI applications on Google Cloud and want standardized, maintained integrations between language models and Google services. It suits projects that rely on Google Cloud infrastructure and need reliable, officially-supported protocol implementations rather than custom integrations. The repository serves as both a reference implementation and a practical starting point for teams already committed to the Google Cloud ecosystem.
The project maintains a curated list of official Google MCP servers with documentation links to each service's implementation details. Open-source server implementations are provided alongside the remote offerings, giving teams options for deployment architecture. The repository includes deployment guidance specific to Google Cloud, helping teams understand how to operationalize these servers in their infrastructure.
