No need to envy Claude Code and Cursor anymore!
Google just released a bombshell product — Gemini CLI, an open-source AI agent that runs in the terminal.
Google has truly disrupted the market this time:
Free to use: Simply log in with your personal Google account to get a free Gemini Code Assist license.
Powerful model: The free version gives you direct access to the Gemini 2.5 Pro model and its massive 1 million token context window.
Generous usage: It offers the industry's largest free quota – 60 requests per minute, 1000 requests per day, absolutely free!
This means that for most individual developers and learners, you can use this powerful AI companion in your terminal almost without restrictions, whether for writing code, debugging, learning, or automating tasks—it's more than enough.
As soon as the news broke, the GitHub project instantly gained 9k+ Stars.
GitHub address:
https://github.com/google-gemini/gemini-cli
Positioning: All-round AI Command Line Assistant
Google understands developers and the importance of the Command Line Interface (CLI). Gemini CLI deeply grasps this, functioning not just as a code generator but as a versatile AI assistant capable of handling various tasks:
Coding and Debugging: Writing code, understanding complex codebases, fixing bugs, generating test cases.
Content Creation and Research: Writing documentation, generating blog posts, conducting in-depth technical research.
Task Management and Automation: Breaking down complex tasks into steps and automating them with scripts.
For example, by linking with Google Veo 3 and Imagen models, Gemini CLI can complete a complex task like “creating a short video of an Australian orange cat adventure story” directly from the command line, including generating video and synthesizing narration.
Core Features
Gemini CLI's powerful built-in tools are based on an open design philosophy:
Google Search Integration: It can fetch real-time web content, providing the model with the latest external information, so answers are no longer limited by outdated training data.
Fully Open-Source and Extensible: Completely open-source under the Apache 2.0 license. You can review the code, verify its security, and even contribute to it. It also supports MCP (Model Context Protocol) and plugin extensions, making it highly versatile.
Highly Customizable: You can customize system prompts and instructions through methods like the GEMINI.md file, training Gemini CLI to be your exclusive, most understanding assistant.
Seamless Integration with Existing Workflows: Supports non-interactive mode, allowing easy invocation in your automation scripts.
Seamless Integration with VS Code
Gemini CLI shares the same core technology with Google's AI coding assistant in VS Code, Gemini Code Assist.
This means that whether you are developing projects in the terminal or in graphical VS Code, you can get the same powerful and consistent “AI agent” experience. Code Assist's “agent mode” can also help you complete complex multi-step tasks such as writing tests, fixing errors, building features, and even migrating code.
Usage
Installation and usage are very simple.
So, everyone, there's no need to subscribe to various model services annually, haha. But frankly, Google's ability to offer these free services primarily stems from using its own chips, which allows for various modifications and optimizations—an advantage OpenAI, Anthropic, and others do not possess.
References:
https://blog.google/technology/developers/introducing-gemini-cli-open-source-ai-agent/