Gemini CLI
Gemini CLI is Google's open-source AI agent under Apache-2.0 that brings Gemini models into the terminal. It signs in with a personal Google account for a free tier the README puts at 1,000 requests a day, or with a Gemini API key or Vertex AI. It has Google Search grounding, file and shell tools, MCP support, a 1 million token context window and a GitHub Action. The code is open; the models are Google's.
What does Gemini CLI do?
"Gemini CLI is an open-source AI agent that brings the power of Gemini directly into your terminal." Its README leads with the free tier: "60 requests/min and 1,000 requests/day with personal Google account", and with a Gemini API key "1000 requests/day with Gemini 3 (mix of flash and pro)". It uses Gemini 3 models with a 1 million token context window; no other provider and no local model is mentioned.
Built-in tools are "Google Search grounding, file operations, shell commands, web fetching." MCP support adds custom integrations through a settings file. It runs non-interactively with gemini -p and JSON or streamed JSON output for scripts, keeps conversation checkpoints, reads GEMINI.md context files, has a VS Code companion, and ships a GitHub Action for pull request review and issue triage that answers @gemini-cli mentions. Three release channels run weekly previews, weekly stable builds and nightlies. Sandboxing has its own documentation page whose contents are not in the README.
Key facts
- Licence: Apache-2.0 with no added conditions.
- Free tier, per the README: 60 requests a minute and 1,000 a day with a personal Google account; 1,000 a day with a Gemini API key; usage-based billing and Vertex AI for enterprises; a paid Gemini Code Assist licence is also recognised.
- Models: Gemini 3, with a 1 million token context window; no other provider or local model is mentioned.
- Built-in tools: Google Search grounding, file operations, shell commands and web fetching; MCP servers configured in ~/.gemini/settings.json.
- Non-interactive mode with gemini -p, --output-format json and stream-json; conversation checkpointing; GEMINI.md context files.
- A VS Code companion and a GitHub Action for PR review and issue triage.
- Install: npx @google/gemini-cli with no install, npm install -g @google/gemini-cli, Homebrew on macOS and Linux, or MacPorts.
- Release channels: preview weekly, stable weekly, nightly daily. Latest tagged release when read: v0.60.0 on 2026-09-15; 107,044 stars on 2026-09-18.
What does it replace, and where does it fall short?
Gemini CLI replaces Claude Code for a developer on Google's models, with a free tier that needs no card and no key, search grounding built in, and Apache-2.0 source. On the open-source alternatives to Claude Code page it is the pick for a free tier on a Google account.
Where it falls short of an open-source replacement: every model call goes to Google, whether through the free tier, an API key or Vertex AI, and no local model is offered, so the facts panel marks it as not running on your own machines. The free tier's limits and terms are Google's to change. For local models in a terminal agent, Crush, Qwen Code, which began as a fork of this project, or gptme.
How does Gemini CLI run?
Run npx @google/gemini-cli with nothing installed, or npm install -g @google/gemini-cli, brew install gemini-cli, or MacPorts. Sign in with a personal Google account for the free tier, or set a Gemini API key from AI Studio, or point at Vertex AI with Google Cloud credentials. Windows is not named in the README's install section, though npm runs there. Google's paid paths are usage-based API billing, a Gemini Code Assist licence, and Vertex AI.
Who is Gemini CLI for?
A developer who wants a capable terminal agent for free on a Google account, with search grounding and a large context window, and who is fine with Google running the model. Anyone who needs the model on their own hardware should look elsewhere on the same page.
What limits does the README state?
From the README: preview releases "will not have been fully vetted and may contain regressions or other outstanding issues", and nightly builds should be assumed to have pending validations and issues.
Questions people ask
Is Gemini CLI open source?
The agent is: Apache-2.0 with no added conditions, and the README says so. The models are Google's hosted Gemini 3 models, reached through a Google account, an API key or Vertex AI; there is no other provider and no local model. This site labels the code open source and marks the tool as not running on your own machines.
Is Gemini CLI free?
There is a free tier, by the README: 60 requests a minute and 1,000 requests a day with a personal Google account, and 1,000 a day with a Gemini API key on a mix of Flash and Pro models. Beyond that, API usage is billed, and enterprises use Vertex AI or a Gemini Code Assist licence. Those terms are Google's and can change.
Does Gemini CLI support MCP and shell commands?
Yes to both. The README lists built-in tools for file operations, shell commands, web fetching and Google Search grounding, and MCP support for custom integrations configured in a settings file. It also runs non-interactively for scripts with JSON output, and as a GitHub Action for pull request review and issue triage.
How does Gemini CLI compare with Claude Code?
Same shape, different vendor: a terminal agent tied to one company's models, with a free tier on Google's side and open source for the agent code. The open-source alternatives to Claude Code page compares it with OpenCode, Codex CLI, Qwen Code, Crush, gptme and goose, including which of them run local models.
What is the relationship between Gemini CLI and Qwen Code?
Qwen Code's README says it was originally based on Google Gemini CLI version 0.8.2 and has developed independently since its own 0.1 release. Qwen Code added support for OpenAI, Anthropic and Qwen APIs and for local models through Ollama or vLLM, which Gemini CLI does not offer.
Sources
- Gemini CLI README and Apache-2.0 LICENSE: github.com/google-gemini/gemini-cli, read 2026-09-17.
Compared with the others
On the open-source alternatives to Claude Code page, Gemini CLI is the pick for a free tier on a google account. Apache-2.0, with 1,000 requests a day on a personal Google account by its README, search grounding, shell and file tools, MCP and a GitHub Action.
Also on that page: OpenCode for a terminal agent with a planning mode, Qwen Code for switching between hosted and local models at runtime, Crush for the widest provider list with local models.
More coding assistants and agents
An MIT terminal coding agent with a full-access build agent and a read-only plan agent switched with the Tab key, installers for every platform, a beta desktop app, and the most-starred repository in this directory.
OpenAI's Apache-2.0 coding agent that runs locally in your terminal, with a desktop app and IDE extensions, signed in with a ChatGPT plan or an OpenAI API key, so the code is open and the models are OpenAI's.
A high-performance code editor from the creators of Atom and Tree-sitter, GPL-3.0-or-later with Apache-2.0 components, with an agent panel, external agents over ACP, edit prediction, MCP tools, sandboxing, and local models through Ollama or llama.cpp.
An MIT self-hosted control centre, called Agent Canvas, that runs the OpenHands agent or Claude Code, Codex, Gemini and any ACP-compatible agent across local, Docker and cloud backends, with automations triggered by schedules and webhooks.
An Apache-2.0 coding agent for VS Code, JetBrains, the terminal and a desktop app that edits files as reviewable diffs, runs commands with your approval, works with any model provider including Ollama, and connects tools through MCP.
An Apache-2.0 general-purpose AI agent from the Agentic AI Foundation at the Linux Foundation, written in Rust, with a desktop app, a CLI and an API, 15+ model providers including Ollama, and 70+ extensions over MCP.
Added September 17, 2026. Every claim above comes from the project's README, LICENSE or model card, read on September 17, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/google-gemini/gemini-cli.
Running one of these inside your own environment, with your own data and your own security rules, is the kind of work Reveneau does. Read how a forward deployed engagement works.