Gemini CLI

Gemini CLI brings the power of Google's Gemini models directly into your terminal, enabling developers to understand code, automate tasks, and build workflows with full local project context. It supports querying and editing large codebases, generating apps from images or PDFs, and running complex multi-step automations—all without leaving the command line. Installed via a single npm command, it works across personal and enterprise environments with flexible authentication options. The tool is open source and actively maintained by the Google Gemini team.

• Codebase understanding — Query and navigate large codebases with natural language to get instant explanations and suggestions.
• App generation — Generate applications from images or PDFs by describing what you want directly in the terminal.
• Agent Skills — Use specialized agents for domain-specific tasks to extend Gemini CLI's capabilities.
• MCP server support — Connect to and use remote agents via the Model Context Protocol for expanded tool integrations.
• Extensions — Browse and install community extensions to add new tools and capabilities to your workflow.
• Checkpointing & Rewind — Automatic session snapshots let you rewind and replay conversations to recover from mistakes.
• Plan mode — Use a safe, read-only planning mode to design complex changes before executing them.
• Headless mode — Run Gemini CLI programmatically or in scripts for full automation pipelines.
• IDE integration — Integrate Gemini CLI with your favorite IDE for a seamless development experience.
• Sandboxing — Isolate tool execution for safer operation in sensitive environments.
• Model routing & selection — Automatic fallback resilience and manual model selection to choose the best Gemini model for each task.
• Token caching — Performance optimization that reduces latency and cost on repeated context.
• Enterprise configuration — Fine-grained policy engine and trusted folder controls for professional environments.