October 15, 2025 — Google has announced a major enhancement to the Gemini CLI, introducing full interactive terminal support that allows developers to run rich, complex shell commands—like vim, top, or even an interactive git rebase -i—directly inside the Gemini CLI environment.

With this release, Gemini CLI v0.9.0 becomes a fully interactive command-line environment, eliminating the need to switch between separate terminals for advanced workflows.

“Developers told us they wanted a more powerful and natural terminal experience,” said the Gemini CLI team at Google. “This update gives you just that—everything from editing to monitoring to interactive git workflows, all within Gemini CLI, without breaking context.”

Context that Stays with You

In earlier versions, developers had to exit Gemini CLI to run interactive commands, which meant losing context and environment state. Now, thanks to new pseudo-terminal (PTY) support, commands that depend on terminal control codes—such as editors, monitors, and interactive tools—can run within Gemini CLI’s own context, seamlessly.

This ensures a continuous and integrated workflow: every command, from system-level to AI-driven, happens in one unified environment.

How It Works: A Virtual Terminal Inside Gemini

Under the hood, Gemini CLI now leverages Node-PTY, spawning each shell command in a virtual terminal (PTY) process. The PTY acts as an intermediary that emulates the behavior of a physical terminal.

A new terminal serializer captures every aspect of the PTY’s state—including text, color codes, and cursor positions—rendering it live in your Gemini interface.

Think of it as a real-time video feed of your shell session: what you see is an active, responsive terminal that you can directly interact with.

Full Two-Way Interactivity

The new Gemini CLI shell isn’t just for display—it’s fully interactive.

Here’s what developers can now do:

• Type directly into running processes — every keystroke is sent to the PTY in real time.

• Resize the terminal dynamically — layouts and interfaces adjust instantly.

• Experience full-color output rendering, matching native terminal behavior.

• Focus instantly on the interactive shell with Ctrl + F.

These improvements bring parity with standalone terminals, letting developers use familiar tools and workflows inside Gemini without compromise.

What You Can Now Run Natively in Gemini CLI

With the interactive shell enabled, developers can now run complex terminal applications, such as:

• Code editors like vim, nvim, or nano

• Interactive REPLs for Python, Node.js, or Go

• Git workflows like git rebase -i or git add -p

• System monitors such as htop and top

• Setup tools like npm init, ng new, or cargo new

• Interactive gcloud commands requiring user input

Each of these now behaves as it would in a native shell—complete with interactive prompts, real-time updates, and terminal responsiveness.

Getting Started

The interactive shell is enabled by default in Gemini CLI v0.9.0.

To upgrade, run:

npm install -g @google/gemini-cli@latest

Once installed, simply launch Gemini CLI and start using your favorite interactive commands inside it—no extra setup needed.