What you need to know
- Gemini CLI is a new, local, and open-source AI agent now available for your command-line interface.
- It’s powered by Gemini 2.5 Pro with its one-million-token context window, but can be tweaked with extensions and more data sources.
- You can install Gemini CLI in preview for free with a personal Google account.
Google is bringing Gemini to a new place that’s particularly handy for developers and programmers: the command-line interface (CLI). The company today announced Gemini CLI, an open-source AI agent that aims to upgrade your terminal while keeping security and locality at the forefront. Gemini CLI is powered by Gemini 2.5 Pro and offers a one-million-token context window — and best of all, it’s free.
While the latest Gemini 2.5 Pro model is at the heart of the Gemini CLI experience, its capabilities don’t end there. The tool can also use Google Search to access real-time data in addition to the model’s local knowledge base. It supports growing open specifications, like the Model Context Protocol (MCP), so users can add their own sources and extensions.
Since Gemini CLI works within your terminal, you can personalize it to match your current workflows and processes. The preview is fully open-source , meaning it can be peer-reviewed to ensure it’s secure. You would probably want to see what you’re adding to your terminal before installing Gemini CLI, and Google is letting you do just that by opening it up.
Google says that Gemini CLI will add “powerful AI capabilities, from code understanding and file manipulation to command execution and dynamic troubleshooting” to your command-line interface. It’ll also make it possible to “write code, debug issues and streamline your workflow with natural language” without using a separate chatbot or IDE, per the company.
Speaking of integrated development environments (IDEs), Gemini CLI is integrated with Gemini Code Assist for when you’d rather work in an app like VSCode. The two tools are sharing tricks, as Gemini Code Assist now gets an agent mode to match the functionality of Gemini CLI. The idea is that whether you want to work in a terminal or an IDE, you’ll have an AI agent to help in Gemini.
Developers and programmers that are more familiar with CLIs than graphical user-interfaces (GUIs) might find it easier to integrate Gemini CLI into their processes than a traditional AI chatbot. Similarly, this terminal add-on could give peace-of-mind to users that want to use local, open-source AI on their devices without the hassle.
It’s available in preview starting today for free with a personal Google account, and you can install it here.
Source link