Google launches Gemini CLI, a command-line AI coding assistant
Google rolled out Gemini CLI, a new agentic AI tool that runs from your terminal. It links the company’s Gemini AI models directly to local codebases. Developers can ask it to explain code, write features, debug, or run commands using natural language.
Gemini CLI aims to embed Google’s AI into developers’ daily workflows. It joins tools like Gemini Code Assist and Jules, Google’s async AI helper. But Gemini CLI targets CLI fans directly, competing with OpenAI’s Codex CLI and Anthropic’s Claude Code – known for speed and easy integration. 
Gemini 2.5 Pro, launched in April, boosted Gemini’s popularity among devs. That popularity helped fuel third-party tools like Cursor and GitHub Copilot. Google wants to cut out middlemen, pushing in-house AI options like Gemini CLI.
Gemini CLI isn’t just for coding. It can create videos with Google’s Veo 3, generate research reports via Deep Research, or pull live data using Google Search. It also hooks into MCP servers for external database access.
Google is open-sourcing Gemini CLI under the permissive Apache 2.0 license, inviting community contributions on GitHub. Free users get 60 model requests per minute and 1,000 per day—roughly double the average current usage.
The launch follows rising doubts over AI coding tools. A 2024 Stack Overflow survey found only 43% of devs trust AI-generated code accuracy. Studies have flagged AI for introducing errors and missing security flaws.
No word yet on how Google will handle that risk inside Gemini CLI. For now, the tool is live, open source, and ready to run from your terminal.
