Google has announced Gemini CLI, an open-source tool designed to bring its AI models directly into the developer's terminal. The Mountain View giant is not the first company to think of putting an AI assistant in the command line. We've seen similar projects from Google's competitors in this space, like OpenAI with Codex CLI and Anthropic with Claude Code (released last February).
What seems to be Google's main advantage here is the usage allowance, which is surprisingly generous for a free offering. Logging in with a personal Google account grants users up to 60 model requests per minute and a total of 1,000 requests per day at no charge. This access is piped directly to Gemini 2.5 Pro, Google's powerful model known for its massive 1 million token context window, giving individual devs a lot of room to experiment and build without hitting a paywall.
If you are a professional developer who needs to run multiple agents at the same time, or if you want to use specific models beyond the default, Google has other options. You can use a Google AI Studio or Vertex AI key for pay-as-you-go billing, or just get a Gemini Code Assist Standard or Enterprise license to unlock more power.
Here's what you can do with the CLI:
- Ground prompts with Google Search so you can fetch web pages and provide real-time, external context to the model
- Extend Gemini CLI's capabilities through built-in support for the Model Context Protocol (MCP) or bundled extensions
- Customize prompts and instructions to tailor Gemini for your specific needs and workflows
- Automate tasks and integrate with existing workflows by invoking Gemini CLI non-interactively within your scripts
Getting it running is straightforward, assuming you are set up for modern web development. To install Gemini CLI, you just have to make sure you have Node.js 18 or later installed first. From there, you can either run it directly with npx, like this:
npx https://github.com/google-gemini/gemini-cli
Or install globally for easier access later:
npm install -g @google/gemini-cli
Because the tool is open source, anyone can dig into the code (hosted on GitHub) to understand exactly how it works or verify its behavior. Powered by Yargs and licensed under Apache 2.0, the tool should feel familiar to anyone who's built CLI apps with Node.js, making it easier to contribute to.
The CLI also shares technology with Google's IDE plugin, Code Assist. This creates a more unified experience for developers using Google's AI tools, whether they are working in VS Code or directly in the terminal. The integration works for users on the free tier and for developers who are subscribed to the Standard or Enterprise plans.
You can find more information in the official Google announcement post.
0 Comments - Add comment