Google’s Gemini CLI just got a massive upgrade — the Extensions feature has unlocked its full potential.
These extensions let you integrate your favorite third-party tools and services directly into your AI workflow. In other words, you can now do things in the terminal — like drawing architecture diagrams or deploying projects — that previously required opening multiple apps.
Let’s dive in.
🚀 The Basics of Gemini CLI
Before we get into extensions, let’s briefly cover what Gemini CLI can do out of the box.
- Built-in Gemini Models – It integrates Google’s Gemini models, offering strong reasoning and generation capabilities.
- Developer Assistance – It can write, debug, and refactor code, or even generate a README based on your project files.
-
Built-in Tools – Gemini CLI can call essential tools by itself, such as:
- Google Search: Fetch the latest web information.
- File Operations: Read or write files in your project.
-
Shell Commands: Execute commands like
lsorgrep.
-
Project Context: You can add a
GEMINI.mdfile in your project to describe background, coding standards, and other details — the AI will tailor its responses accordingly.
Gemini CLI’s free plan supports 60 requests per minute and 1,000 requests per day. The Gemini 2.5 Pro version also offers a 1 million token context window, meaning you can feed an entire project’s codebase into it for deep context understanding.
⚙️ How to Install Gemini CLI
Gemini CLI requires Node.js 20 or above — this is also necessary for installing extensions later.
You can use ServBay to quickly set up the required Node.js environment and other development tools with one click.
- Download and install ServBay.
- In the left menu, go to Packages, select Node.js, and click to install your desired version. (Make sure to choose version 20 or higher.)
- Once installed, open the terminal and verify the setup with
node -v.
✅ Now your environment is ready.
To install Gemini CLI, just run:
npx https://github.com/google-gemini/gemini-cli
Then follow the prompts:
- Pick your preferred theme color.
- Sign in with your Google account.
You’re done — the entire process takes less than 5 minutes.
🔌 What Are Extensions?
Extensions are plugins for Gemini CLI. Installing one teaches the CLI how to use a new external tool.
For example, by installing a diagram tool extension, you can simply say:
“Generate a user login architecture diagram.”
…and Gemini CLI will call that tool and complete the task directly from your terminal — no app-switching needed.
You can use public extensions or create your own. If your team has internal tools or a fixed automation pipeline (e.g., packaging, deployment, notifications), you can bundle them into an extension so anyone on your team can trigger it with one command.
🧩 Example: Creating a Diagram with the Nano Banana Extension
Let’s go through a practical example using Nano Banana, one of the most popular extensions recently added to Gemini CLI.
Step 1: Find and Install the Extension
Go to the Gemini CLI Extension Gallery, search for Nano Banana, and follow the GitHub link.
Before installing, ensure that:
- You have Node.js 20+
- npm is available (it’s installed automatically with Node.js if you used ServBay)
Install the extension by running:
gemini extensions install https://github.com/gemini-cli-extensions/nanobanana
Step 2: Configure and Verify
- Set Your API Key: Nano Banana uses an API key — you can reuse your existing Gemini API key.
export NANOBANANA_GEMINI_API_KEY="your_gemini_api_key"
- Restart Gemini CLI and check whether the extension was successfully added:
/extensions list
If you see nanobanana listed, everything is ready to go.
Step 3: Generate Your First Architecture Diagram
Use the /diagram command:
/diagram "Generate a technical architecture diagram for an AI agent." --type="architecture" --complexity="simple"
Approve the permissions when prompted, and wait a few seconds — a fully generated AI architecture diagram will appear.
Step 4: Refine Using Natural Language
If the first result looks messy, just ask Gemini CLI to fix it:
“Make it more professional, aligned, and consistent in font style.”
It will understand and regenerate a cleaner version — you can iterate this process until you’re satisfied.
🧠 Final Thoughts
The Gemini CLI’s extension system turns it from a simple AI chat tool into a powerful command-line hub that can integrate dozens of developer workflows — from diagramming to deployment.
And with ServBay, setting up your Node.js 20+ environment and other tools becomes effortless — it’s the easiest way to deploy your dev environment with one click.
Gemini CLI + ServBay = A developer’s dream workflow.
Try it out and see how seamless your terminal life can become.











Top comments (0)