How to Get Started with Google’s Open‑Source Gemini CLI – A Hands‑On Guide

Google has open‑sourced Gemini CLI, a powerful terminal AI programming assistant that provides the industry’s highest free quota: 60 model requests per minute and 1,000 requests per day using the Gemini 2.5 Pro model.

Within two days of release the project attracted over 31,000 stars, sparking great interest among developers.

Key features include:

Context window : supports up to 1 million tokens.

Multimodal support : can generate applications from UI sketches or PDF documents.

Automation capabilities : can automatically query pull requests and handle complex code merges.

MCP extensions : integrates various tools via the MCP protocol to enhance coding.

Google Search integration : provides real‑time, reliable information lookup.

Prerequisites for installing Gemini CLI are Node.js 18 or newer, at least 4 GB of RAM (8 GB recommended), and 500 MB of free disk space. npm install -g @google/gemini-cli After installation, run the command gemini to launch the interactive interface. The tool will prompt you to select a theme, log in with a Google account, and authorize access via the default browser.

Once authorized, you can start sending prompts. For example, typing “你好！请介绍一下你自己” receives an immediate response, confirming the CLI is functional.

The article demonstrates building a simple web application that helps high‑school students choose university majors based on their exam scores. The prompt sent to Gemini CLI was:

帮我创建一个网页应用，需要实现的功能是：当用户输入高考的分数时，可以帮用户筛选出可以报考的院校，需包含院校、排名、批次、性质、历史录取最低分这些信息。

Gemini generated a detailed implementation plan, installed required dependencies, and produced the code. After the build completed, the app could be previewed in Chrome, allowing users to input a score and see a list of eligible universities with rankings, batch, nature, and minimum admission scores.

To further enhance the app, the following prompt was used to categorize universities:

以用户输入的分数 x 为参考，按如下规则筛选并分类院校：</code><code>- 冲：高于 x 15 分内的院校</code><code>- 稳：在 x 以下 20 分的院校</code><code>- 保：在 x 以下 30 分的院校</code><code>在表格的最左侧增加 “属性” 一列，分别显示 “冲”，“稳”、“保”，同时分别用 “黄”、“蓝”、“绿”代表它们的每一行背景色。

After sending the prompt, Gemini quickly updated the code, added the new column, and applied the color coding. The final interface displays the filtered universities with visual cues for “冲” (high chance), “稳” (stable), and “保” (safe) categories.

Overall, Gemini CLI offers a free, open‑source AI coding experience, though its interaction flow is less polished than commercial AI editors. Continued community contributions are expected to improve usability.

The release of Gemini CLI as a free, open‑source tool adds significant value for developers and may reshape the AI programming tool market.