Gemini-CLI Masterclass | The Only Guide You Need | Gemini 2025

Video Description

The ultimate, start-to-finish guide to Gemini CLI, follow along as we install, authenticate, configure settings.json, use built-in tools, and run Gemini non-interactively. Perfect for beginners and devs who want a single, complete reference. ## What’s inside: - Fast install (Cloud Shell) - Auth (OAuth options) - Settings.json - Shell mode - Custom commands - Running commands, non-interactive mode, - MCP setup ## Official resources to follow along: - Codelab for Hands-on with Gemini CLI: https://codelabs.developers.google.com/gemini-cli-hands-on?hl=en#0 - Gemini CLI repo: https://github.com/google-gemini/gemini-cli ## Connect with me: 🐦 Twitter/X: https://x.com/shishir_who 📸 Instagram: https://instagram.com/programmatic.ly 📝 Tech Blog: https://hashnode.com/@ShishirSrivastav 💻 GitHub: https://github.com/Shishir420-GIT 💼 LinkedIn: https://linkedin.com/comm/mynetwork/d... 🎯 Book a Call: https://topmate.io/shishir_srivastav Subscribe for more AI & tech content! If this helped, smash Like, Subscribe, and drop a comment with what you want next? demo automations, GitHub Actions, or advanced tool usage? Timestamps: 00:00 - Intro & What is Gemini CLI? 01:00 - Free tier explained (limits & token usage) 02:20 - Tools overview: Google Search vs Web Fetch 03:30 - Starting Google Cloud Lab environment 04:30 - Creating a new project in GCP 05:30 - Setting up billing & free credits 06:30 - Launching Google Cloud Shell 07:10 - Starting Gemini CLI for the first time 08:10 - Fixing "Gemini command not found" error (npm install) 09:30 - Creating a project folder for Gemini CLI 10:30 - Running Gemini inside project folder 11:30 - Authentication: Google login, API Key, Vertex AI 13:30 - Setting Project ID in gcloud 14:30 - Understanding `settings.json` file 16:30 - User-level vs Project-level configs 18:00 - Example of MCP server config in settings.json 19:10 - First Gemini CLI interaction (Google Search) 21:00 - Commands & shortcuts (/quit, /m, etc) 22:00 - Changing models in Gemini CLI 23:00 - Non-interactive mode with `-p` flag 24:20 - Handling errors (quota exhausted, fallbacks) 25:00 - Using YOLO flag (auto-accept commands) 26:20 - Built-in tools: edit, read, write, search 27:20 - Viewing available MCP servers 28:10 - Example: Search + Write results to file 29:20 - Reading files with Gemini CLI 30:00 - Shell mode (! commands inside Gemini) 31:20 - What is MCP protocol & why it matters 33:00 - Adding custom MCP servers in settings.json 35:02 - Creating a GitHub Personal Access Token (PAT) 35:47 - Token permissions, expiry & repo access setup 36:34 - Generating & pasting the PAT into `settings.json` 37:06 - Validating token & reloading Gemini CLI 38:10 - Testing GitHub MCP connection with `/mcp` 38:58 - Verifying GitHub MCP: fetching username & profile 39:31 - Why & when to use MCP servers (GitHub, Firebase, Workspace) 40:32 - MCP toolbox example (databases, Google Workspace MCPs) 41:05 - Customizing the `Gemini.md` file (personas, memory, context) 42:44 - Creating and editing a new `Gemini.md` file 44:40 - Restarting Gemini to load new memory configs 45:34 - Creating custom commands in Gemini CLI 46:45 - Example: Adding `/refactor` command 48:19 - Testing & verifying custom command setup 50:10 - Recap: Project setup, tools, prompts, shell commands, MCPs, memory, and custom commands 52:00 - Use case: “Pair Coding with Gemini CLI” demo prompt 53:08 - Final thoughts, resources & next steps