Google is bringing its proprietary AI more directly into coder workflows.
On Wednesday, the company announced Gemini CLI, a free new agentic AI tool that integrates directly with a command line interface (CLI). Google positioned the agent as an immediate link between coders and Gemini 2.5 Pro, the latest iteration of Google’s flagship AI model, saying it “provides lightweight access to Gemini, giving you the most direct path from your prompt to our model” in a blog post.
Also: The best AI for coding in 2025 (including a new winner – and what not to use)
Developers can use Gemini CLI for writing and debugging code via natural language prompts, generating content, deep research, task management, and more. First introduced with an experimental version in March, Gemini 2.5 Pro is a reasoning model, meaning it’s designed to execute more complex queries step-by-step, including coding. Google recently expanded the model’s capabilities by introducing a Deep Think mode and doubling the limit of daily prompts for users.
Gemini CLI is also integrated with Gemini Code Assist, Google’s AI coding assistant, meaning users across free, Standard, and Enterprise plans can access it in VS Code.
<!–>
In a similar spirit, Google is now trying to promote usage of its new Gemini CLI tool by allowing up to 60 model requests per minute and 1,000 per day, completely free – “the industry’s largest allowance,” according to the blog post. The company says it has also open-sourced the new tool via Apache 2.0, one of the software development industry’s most permissive licenses.
Gemini CLI marks Google’s latest effort to incorporate its proprietary AI tools directly into coders’ day-to-day workflows. At its annual developer conference I/O last month, the company also announced the public beta launch of Jules, its asynchronous AI coding assistant.
Also: Google’s Jules AI coding agent built a new feature I could actually ship – while I made coffee
Developers are often the first-considered demographic in the ongoing generative AI race; Gemini CLI is the latest in a slew of developer AI tools intended to automate more basic tasks so that coders can focus on higher-demand work. In April, Anthropic launched Claude Code, another agentic AI tool that makes the chatbot accessible directly via terminals. Open-source tools are also increasingly relevant to a job market made uncertain by AI’s rapidly evolving capabilities.
Want more stories about AI? Sign up for Innovation, our weekly newsletter.
AI coding assistants have seen widespread adoption in recent years. A 2024 survey of web developers from the Q&A site Stack Overflow found that 76% of respondents were actively using or planned to start using AI tools in the coding process, with increased productivity being the most widely cited benefit. That same survey, however, also found widespread distrust of the accuracy and reliability of AI coding assistants.
Also: How AI coding agents could destroy open source software
Gemini CLI is now available as a preview. To access, you can install it for free through GitHub.
Google says it encourages users to provide feedback via the company’s GitHub repository. “We fully expect (and welcome!) a global community of developers to contribute to this project by reporting bugs, suggesting features, continuously improving security practices and submitting code improvements,” the company wrote in the blog post.
