Gemini CLI is Google's open-source command-line AI agent that brings Gemini 2.5 Pro directly into the terminal. Released in mid-2025 under an Apache 2.0 license, it positions itself as a free, transparent alternative to closed-source coding agents such as Claude Code and Cursor, and is distributed via the official google-gemini/gemini-cli GitHub repository.
Under the hood, Gemini CLI uses a ReAct (reason-and-act) loop that combines Gemini 2.5 Pro's 1-million-token context window with built-in tools and Model Context Protocol (MCP) servers. It can read and edit files, run shell commands, search the web, and orchestrate multi-step tasks such as bug fixes, feature scaffolding, and test generation. The CLI also integrates with Gemini Code Assist for Google Cloud users who need IAM, audit logs, and Vertex AI key management.
Because Gemini CLI is a Google product rather than a venture-backed startup, there is no traditional funding history; total external funding is effectively zero and development is funded internally by Google. The project was first announced on the Google Developers blog in June 2025 and is maintained by Google's developer tools team alongside contributions from the open-source community.
Developers can sign in with a personal Google account to access Gemini 2.5 Pro at up to 60 requests per minute and 1,000 requests per day at no cost, an unusually generous quota for a frontier model. Paid usage is available via Google AI Studio API keys or Vertex AI for higher rate limits and enterprise governance, making it suitable for both hobbyists and regulated teams.
Compared with Claude Code, Gemini CLI's main differentiators are its open-source license, the size of its free tier, and direct access to Gemini 2.5 Pro's long context. Early developer feedback notes that Claude Code currently has a more polished agentic experience, but Gemini CLI's pricing and openness make it a strong default for cost-sensitive teams and contributors who want to inspect or fork the agent itself.