Setting Up Walter MCP in Gemini CLI and Workflows

Gemini CLI can read and write files within your project, making it an ideal environment for writing tools alongside coding utilities. By connecting Walter Writes AI's Model Context Protocol (MCP) server, Gemini can run AI detection on markdown files, humanize specific passages that require it, and write the final output directly back to disk without requiring you to manually copy text between a terminal window and a browser tab.

Before proceeding with the setup, a notable change this year affects which part of this guide you should follow. On June 18, 2026, Google discontinued Gemini CLI request support for free-tier users as well as Google AI Pro and Ultra subscribers, transitioning those users to the new Antigravity CLI. Gemini CLI remains fully functional for users with paid Gemini API keys, Gemini Code Assist Standard and Enterprise licenses, and Google Cloud access. If you fall into that category, sections 2 through 8 apply to you. If you were migrated to Antigravity CLI, skip ahead to section 9. The Walter server and overall workflows remain identical; only the configuration file differs.

Stock photo for illustration only, not from the actual event

Walter's MCP server operates as a remote server hosted at https://mcp-server.walterwrites.ai/mcp. It utilizes OAuth authentication, eliminating the need to paste raw API keys. You simply sign in to your Walter account once through a browser window, after which tool calls execute under your Walter workspace subscription plan and usage limits, following the official Gemini CLI integration guide.