When you think of artificial intelligence, you often imagine vocal assistants or well -licked graphic interfaces. However, a large part of the work of the developers still goes through a tool that the general public uses little: the terminal or online command line (CLI in English). It is a black screen in which we type keyboard instructions, without graphical interface, a bit like in the era of MS-DOS.
Free, open source and without bank card
Despite its austere appearance, the terminal remains a very powerful tool for developers. And this is precisely where Google wants to implant its AI assistant: Gemini Cli makes it possible to send instructions in natural language, as we would in a cat, and to obtain useful answers to code faster, automate certain tasks or understand a piece of code.
One of the main advantages of Gemini Cli is its accessibility. The tool is free for the vast majority of users: Google authorizes up to 1,000 requests per day, far above the usual needs of a developer. No need to enter a credit card or manage an API key. Just have a Google account.
The other peculiarity is that the Gemini Cli code is free (Apache 2.0 license). Unlike other solutions such as Codex (OpenAi) or Claude (Anthropic), it is possible to consult the operation of the program, modify it, or even integrate it into its own tools.
Google has also treated security: each order generated by AI requires confirmation before being executed, with options like “authorize once” or “always authorize”. The tool can be carried out in isolated environments (via Docker or the protections integrated into the operating system), and all traffic can be filtered to avoid data leaks.
Be careful however: the model of IA Gemini Pro 2.5 does not turn locally, it is hosted on the Google servers. This means that some information will have to pass through the Internet. But it is you who decide what you share with AI with each request.
GEMINI CLI is not limited to answering questions: he can also adapt to different projects thanks to a modular architecture. Google has set up a system based on a protocol called Model Context Protocol (MCP), which allows the tool to be connected to other services (such as image or videos generators, for example).
If Gemini CLI is free for individual developers, companies will have to pay for certain advanced uses: parallel use on several positions, specific data or security requirements … These services go through professional tools like Green AI, the Google Cloud platform.
But for many, Gemini CLI can serve as a simple and powerful entry point to AI, without commitment or hidden costs. As Ryan J. Salva, product manager at Google summarizes: ” No matter whether you are a student, freelance or employee of a large company, you must have access to the same tools. »
🟣 To not miss any news on the Geek newspaper, subscribe to Google News and on our WhatsApp. And if you love us, .
