Gemini CLI - Agentic coding a Firebase App | Firebase After Hours #18

youtube

Join us for a first look at the future of development: agentic coding. In this stream, we explore how the new Gemini CLI, combined with the Firebase MCP Server and the Firebase Extension, supercharges your AI agent. See how it gets the tools and context it needs to *actually* understand, build, and manage a Firebase app right from your terminal. 🚀 In this stream 01:27 - Welcome & Introduction 05:01 - Icebreaker: Favorite Desserts 07:50 - What is the Gemini CLI? 08:40 - Why use the Command Line for AI? 11:18 - Live Demo: Starting a React App 13:52 - Gemini CLI Extension for Firebase 14:43 - How the Firebase Extension Works (MCP Server) 16:18 - Demo: `firebase init` with AI Logic 22:13 - Demo: Building the Halloween Costume App 25:12 - Demo: First Run & Debugging 26:01 - Demo: Asking Gemini to Fix the Bug 28:06 - Demo: It Works! (Banana Costume) 28:51 - Demo: Adding a Spinner 29:15 - Other Firebase Integrations (Firestore, Auth) 30:10 - Using MCP Server for Admin Tasks 34:32 - "YOLO Mode" (`--yolo` flag) 37:33 - Importance of planning with Gemini 39:46 - Tip: Asking Gemini to interview you 40:57 - How to undo changes (Checkpoints) 45:42 - What is "Jules"? (Autonomous Coding Agent) 50:35 - Q&A: Why isn't Gemini offline? 53:04 - Q&A: Dealing with inconsistency 56:07 - Q&A: Managing API credit costs 01:03:33 - How to file feature requests (User Voice) 01:06:31 - Wrap-up #Firebase #Gemini #GeminiCLI #AgenticCoding #AI #ArtificialIntelligence #AppDevelopment #Developer #React #Goo