Google is launching Gemini CLI, a new tool designed to bring its Gemini AI models directly into developers’ local coding environments.
Announced today by Google, Gemini CLI introduces a powerful agentic AI system that runs from a terminal and allows programmers to interact with their codebases using natural language.
According to Google, the new tool enables developers to perform a range of tasks, including explaining complex code, generating new features, debugging, running shell commands, all through conversational prompts.
“It provides lightweight access to Gemini, giving you the most direct path from your prompt to our model. While it excels at coding, we built Gemini CLI to do so much more. It’s a versatile, local utility you can use for a wide range of tasks, from content generation and problem solving to deep research and task management,” Google said in a blog post.
A Push to Own the AI Dev Experience
The release of Gemini CLI also opens up access to other Google AI models, including Veo 3 for video creation and the Deep Research agent for compiling reports, creating a flexible interface not just for software engineers but for researchers and technical creators across fields.
“We’ve also integrated Gemini CLI with Google’s AI coding assistant, Gemini Code Assist, so that all developers — on free, Standard, and Enterprise Code Assist plans — get prompt-driven, AI-first coding in both VS Code and Gemini CLI,” Google added.
More Than Just Code
In line with modern developer culture, Google is open-sourcing Gemini CLI under the Apache 2.0 license, a permissive framework that allows for broad adoption and community-led innovation.
The company expects an active ecosystem to grow around the tool on GitHub, where developers can contribute new features, improve reliability, and share use cases.
To further encourage adoption, Google has also introduced generous usage thresholds, including free-tier, where users can make up to 60 model requests per minute and 1,000 per day.
What This Mean for Users
- Developers can now generate, debug, and explain code directly from their terminal using natural language, speeding up everyday tasks.
- Gemini CLI runs locally, giving users more control over their codebase without needing to rely on cloud-only tools.
- Beyond coding, users can generate videos, compile research, and access real-time data—all from the same interface.
- With an Apache 2.0 licence, developers can customise, extend, and contribute to the tool, shaping its evolution.
- Up to 1,000 requests per day are allowed for free, enabling extensive use without upfront costs.
A New Era of AI-Powered Development
Since the launch of Gemini 2.5 Pro in April 2025, Google’s AI models have gained significant traction within the developer community.
Experts note that the model’s ability to handle complex coding tasks with high accuracy has fuelled its widespread adoption through third-party platforms like Cursor and GitHub Copilot.
Now, with the introduction of Gemini CLI, Google is moving to bring those developer interactions in-house, offering a first-party solution designed to compete directly with existing external tools.
