Google Unveils Gemini CLI for Developers’ Command-Line Needs

Mountain View, CA — Google has launched Gemini CLI, an open-source AI tool designed for developers to use directly within their command-line interface (CLI). Announced on June 25, 2025, Gemini CLI aims to enhance coding efficiency by integrating Google’s Gemini AI models into existing workflows.

Gemini CLI offers developers streamlined access to AI capabilities such as coding assistance, problem-solving, and research generation. It features a one-million-token context window, ensuring developers can engage comfortably with complex coding tasks. “It provides lightweight access to Gemini, giving you the most direct path from your prompt to our model,” a Google spokesperson stated.

The tool supports natural language requests, allowing users to ask Gemini CLI to explain code segments, write new features, debug, or execute commands. By integrating with Google’s existing AI coding assistant, Gemini Code Assist, it ensures that developers can benefit from AI capabilities across multiple platforms, including IDEs like Visual Studio Code.

For those looking to utilize Gemini CLI, signing up for a free Gemini Code Assist license via a personal Google account is all it takes. This free license provides users with access to 60 model requests per minute and 1,000 requests per day—offering what Google claims is the industry’s largest allowance.

Developers interested in trying Gemini CLI can preview it now, with the added benefits of extensibility via the Model Context Protocol (MCP), enabling customization to fit personal workflows. Additionally, since Gemini CLI is open-source, developers are encouraged to inspect and contribute to its ongoing development through GitHub.

While Google has provided a robust free offering encouraging adoption, details on usage beyond the free limits remain unclear. As AI coding tools continue to gain popularity, Google aims to build strong relationships with developers through Gemini CLI, positioning it as a competitive option among existing AI coding utilities.