Gemini Command Line Tool Tips and Tricks

Gemini CLI – Pro Tips & Complete Usage Guide

This guide is also available to star/follow on our GitHub repository.

---

Introduction

Gemini CLI is an open-source AI assistant that brings the power of Google’s Gemini model directly into your terminal. It works as a conversational, agentic command-line tool — meaning it can:

  • Understand your requests
  • Choose and execute the right tools (shell commands, file edits, etc.)
  • Carry out multi-step plans to improve your development workflow

In practice, Gemini CLI acts like a highly capable pair programmer and command-line assistant. It excels at:

  • Coding tasks
  • Debugging
  • Generating content
  • Automating system operations

… all via natural language prompts.

---

Gemini CLI Screenshot

Example:

gemini "Write a Python script to scrape headlines from a news site"

Gemini CLI will generate the solution and can execute it — making it far more than just a static code generator.

---

Getting Started

Installation

You can install Gemini CLI globally via npm:

npm install -g @google/gemini-cli

Or run it without installation using `npx`:

npx @google/gemini-cli

Gemini CLI is built with Node.js/TypeScript and runs on all major platforms.

---

Authentication

On first launch, you’ll need to authenticate. Two main options:

  • Google Account Login (Free Tier)
  • Use Gemini 2.5 Pro for free with generous limits (≈ 60 req/min, 1,000 req/day).
  • Sign in with your Google account — no billing required.
  • API Key (Paid / Higher-tier Access)
  • Obtain from Google AI Studio
  • Set it as an environment variable:
     export GEMINI_API_KEY="your_api_key_here"

🔹 Note: API key users may get higher quotas and enterprise data protections.

---

Basic Usage

  • Interactive Mode
  • Run `gemini` with no arguments:
  • Gemini can then create files, install dependencies, run tests, etc.
  • One-shot Mode
  gemini -p "Summarize the main points of the attached file. @./report.txt"
  $ gemini
  gemini> Create a React recipe management app using SQLite
  • Piping Input
  echo "Count to 10" | gemini

---

CLI Interface Highlights

  • Slash Commands (/) → control session, tools, settings
  • Bang Commands (!) → execute shell commands directly
  • Safe Mode → requires confirmation before making changes

---

Power User Tips

Below are selected pro tips (this guide contains 28+). Each expands Gemini CLI’s functionality.

---

Tip 1: Persistent Context via `GEMINI.md`

Stop repeating yourself — store project-specific context in `.gemini/GEMINI.md` so Gemini always remembers styles, architecture, or rules.

Example:

# Project Phoenix

## Coding Style
- ESLint + Prettier
- async/await > Promises

## Architecture
- Next.js + Tailwind + Prisma
- JWT auth on all API calls

Use `/memory show` to view current context; `/memory refresh` to reload after edits.

---

Tip 2: Custom Slash Commands

Automate repetitive tasks via `.gemini/commands/*.toml` configs.

E.g., `/test:gen` to auto-generate Jest unit tests from descriptions.

description = "Generate a Jest unit test from requirement"
prompt = """
You are a test engineer. Write a Jest test based on:
Requirement: {{args}}
"""

---

Tip 3: Extend with MCP Servers

Add Model Context Protocol servers for custom tools.

E.g., Figma MCP for design-to-code, Google Docs MCP for document reads.

gemini mcp add myserver --command "python3 my_mcp_server.py" --port 8080

List MCP tools: `/mcp`

---

Tip 4: Memory Add & Recall

Store facts or configs across sessions:

/memory add "Staging RabbitMQ port is 5673"

View: `/memory show`

Refresh: `/memory refresh`

---

Tip 5: Checkpointing & Restore

Snapshot files before tool changes:

gemini --checkpointing

Restore:

/restore list
/restore 

---

Tip 6: Read Google Docs & Sheets

With Workspace MCP integration, reference live docs:

@https://docs.google.com/document/d/

Gemini will pull content into your prompt context.

---

Tip 7: Reference Files & Images (`@` syntax)

Attach exact content to prompts:

Explain this code: @./src/main.js

Works for files, directories, and images.

---

Tip 8: On-the-Fly Tool Creation

Prompt Gemini to write helper scripts during a session, e.g., batch log analyzer.

---

Tip 9: System Troubleshooting

Run in home dir to edit dotfiles (`@~/.bashrc`), check configs, or install tools.

---

Tip 10: YOLO Mode

Auto-approve tool actions:

gemini --yolo

⚠️ Use with caution; skips confirmation prompts.

---

Tip 11: Headless/Scripting Mode

Run single prompts in scripts:

gemini -p "Generate a blog outline"

Custom system prompt:

export GEMINI_SYSTEM_MD="/path/to/system.md"

---

Tip 12: Save & Resume Chats

/chat save 
/chat resume 

Continue multi-day conversations seamlessly.

---

Tip 13: Multi-Directory Workspace

Include multiple folders in context:

gemini --include-directories ./frontend ./backend

---

Tip 14: Organize Files

AI-assisted folder cleanup, categorization, and renaming using image content.

---

Tip 15: Compress Long Conversations

Maintain relevance and stay within context limits:

/compress

---

Tip 16: Shell Passthrough (`!`)

Run shell commands inline:

!ls -lh

Persistent shell mode: `!` (empty) then commands.

---

Tip 17: Every CLI Tool is a Gemini Tool

Any tool in `$PATH` can be used — install helpful CLIs, restrict risky ones.

---

Tip 18: Multimodal Input

Reference images/PDF/audio for analysis with model’s vision capabilities.

---

Tip 19: Customize `$PATH`

Run Gemini with restricted tool availability:

PATH=/usr/bin gemini

---

Tip 20: Token Caching & Stats

Reduce costs via API key caching; check usage:

/stats

---

Tip 21: `/copy` to Clipboard

Copies last output directly to system clipboard.

---

Tip 22: Ctrl+C Mastery

  • Once: abort current action
  • Twice: exit CLI

---

Tip 23: Settings via `settings.json`

Change theme, tool permissions, directories:

{
  "theme": "dark",
  "sandbox": "docker"
}

---

Tip 24: IDE Integration (VS Code)

Gemini can track open files/selection and show AI diffs in VS Code.

---

Tip 25: GitHub Action Integration

Run Gemini CLI as a repo bot for PR reviews, issue triage.

---

Tip 26: Telemetry

Enable OpenTelemetry for usage stats and performance monitoring.

---

Tip 27: Follow the Roadmap

Upcoming: background agents for async tasks; engage via GitHub issues.

---

Tip 28: Extensions

Install modular capabilities:

gemini extensions install 

Examples: Cloud Run deployment, Figma design tools, payment APIs.

---

Easter Egg: Corgi Mode 🐕

`/corgi` toggles animated ASCII corgi across your terminal.

---

Conclusion

Gemini CLI is a Swiss army knife of AI capabilities in your terminal — from persistent memory, multimodal input, custom tooling, automation, and integration with IDEs, CI/CD, and external APIs.

Pairing Gemini CLI with content ecosystems like AiToEarn lets creators generate → publish → monetize across platforms including:

  • Douyin, Kwai, WeChat, Bilibili, Xiaohongshu
  • Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, X (Twitter)

… consolidating AI-assisted development and cross-platform content strategies.

---

Learn more:

---

Would you like me to prepare a condensed quick-reference cheat sheet summarizing these 28 tips for faster lookup? It would be a one-page markdown table.

Read more

Translate the following blog post title into English, concise and natural. Return plain text only without quotes.

ChatGPT Atlas 发布,AI 浏览器大乱斗...

Translate the following blog post title into English, concise and natural. Return plain text only without quotes. ChatGPT Atlas 发布,AI 浏览器大乱斗...

# AI Browsers: When LLM Companies Step In 原创 lencx · 2025-10-22 07:00 · 上海 --- ## Overview Large Language Model (LLM) companies are making moves into the **AI browser** space. From new entrants like **Dia**[1], **Comet**[2], and **ChatGPT Atlas**[3], to established browsers like **Chrome** and **Edge** (which now feature

By Honghao Wang