Google Cloud demos Gemini CLI
Google Cloud demoed a Gemini CLI plus a Model Context Protocol to simplify deployment and migration of agentic apps — a move aimed at making multi‑model, agent‑driven systems easier for enterprises to run. The demo underscores how cloud vendors are productizing agent infrastructure. (x.com)
Google Cloud said it is shipping fully‑managed, remote Model Context Protocol (MCP) servers so agents and MCP clients like the Gemini CLI can point to a single, enterprise‑grade endpoint and leverage existing Google and Google Cloud services. (cloud.google.com) The Gemini CLI reference lists MCP‑exposed commands such as /memory, /stats, /tools and /mcp and documents built‑in tools (grep, terminal, file read/write) that MCP servers can surface to agent workflows. (docs.cloud.google.com) Google Cloud has added gRPC as a native transport option while the MCP specification defaults to JSON‑RPC, positioning MCP for lower‑latency, enterprise networking and observability stacks. (cloud.google.com) A recorded demo from Google Cloud walks through migrating and deploying a full‑stack application using Gemini CLI and chaining multiple MCP servers so the agent behaves like an “intelligent DevOps engineer.” (youtube.com) Google first released Gemini CLI as an open‑source developer tool on June 25, 2025, and the project’s positioning has since emphasized extensibility through MCP to integrate terminal workflows with cloud services. (blog.google) Google’s published codelabs and a short Skills course provide hands‑on tutorials for MCP and Gemini CLI app workflows, while community GitHub tutorials and example MCP server repositories show practical patterns for connecting GitHub, Jira, and other enterprise systems. (codelabs.developers.google.com ) (github.com)