The Gemini CLI, a robust command-line interface facilitating direct interaction with advanced Gemini models like Gemini 2.5 Pro for AI-assisted coding, automation, and debugging workflows, has received a significant 3.0 upgrade. This update introduces a suite of powerful features designed to enhance its capabilities as an AI coding agent, notably addressing tool-calling limitations and fostering a highly customizable development environment, all while remaining completely free for up to 1,000 daily requests.
The core of this transformative update is the introduction of the Gemini CLI Extension framework, also known as GenKit Extension, establishing an open AI extension marketplace. This innovative system enables developers to fully customize the CLI, seamlessly integrating their preferred tools and services directly within the terminal environment, thereby eliminating the need for context switching.
- ✨ Open Ecosystem: Users can install extensions from a growing public repository or any GitHub URL/local path using the
gemini extension installcommand. This marketplace already features top-tier integrations ranging from payment providers to database handlers. - 📦 Lightweight Packaging: Each extension bundles Multi-Content Prompts (MCPs), context files, and custom commands into a cohesive, lightweight package. This design inherently teaches Gemini how to leverage external tools, from local scripts to complex APIs, effectively overcoming previous limitations in tool invocation.
- 🛠️ Customization and Development: Developers can build their own extensions using pre-built templates or from scratch, offering extensive customization. This fosters a community-driven ecosystem where third-party developers contribute diverse capabilities, exemplified by extensions for image generation (Nana Banana) or streamlined API testing (Postman integration).
- 🔄 Workflow Management: Extensions are designed for effortless installation, instant use, and easy management, ensuring a clean and flexible development setup.
Further enhancing the agentic capabilities, the update significantly refines Custom Slash Commands. Initially introduced to define reusable prompts and streamline repetitive tasks, these commands now possess heightened power.
- 🚀 AI Agent Empowerment: Custom slash commands can now invoke the newly released extensions and empower AI agents to translate complex, multi-step instructions into simple, single-line commands.
- 🎯 Improved Tool Calling: Crucially, this enhancement drastically improves Gemini's ability to follow complex tool-calling instructions, a previous weakness, thereby bolstering speed, consistency, and overall productivity in AI-driven workflows.
To support this expanded functionality, a comprehensive documentation portal has been launched, providing thorough guides for getting started and maximizing the utility of every feature. Additionally, continuous refinement is evident through a detailed change log, showcasing ongoing bug fixes and minor edits that contribute to making Gemini CLI a leading terminal agent.
To get started with the upgraded Gemini CLI:
- Ensure Node.js 18 or above is installed on your system.
- Install the CLI globally via npm:
npm install -g @google-gemini/cli. - Launch the CLI by typing
geminiin your terminal. - Authenticate with your Google account (a free process).
Upon activation, users can access powerful models like Gemini 2.5 Pro and explore the new features. The /extensions command allows for managing and listing installed extensions, while other new commands facilitate better interaction with various tools and plugins, such as the Chrome DevTools MCP, which enables AI agents to interact with web browsers for tasks like navigation.