Gemini CLI Configuration Guide

Gemini CLI Configuration Guide

Gemini CLI is Google’s official command-line tool, designed to give developers a powerful AI-assisted coding experience. It supports intelligent code generation, code analysis, debugging assistance, and more.

:rocket: Key Features

• Intelligent code generation - Quickly generate high-quality code
• Code analysis - Deeply understand and analyze code structure
• Debugging assistant - Smartly detect and fix code issues
• Documentation generation - Automatically generate code documentation
• Command-line integration - Seamlessly integrate into your development workflow

:star: Which models can be used with Gemini CLI

All Gemini models on this site are supported in Gemini CLI

:package: Installation Steps

1. Install Gemini CLI

npm install -g @google/gemini-cli  # macOS/Linux需sudo权限

2. Mac & Liunx configuration

Terminal setup method

Run the following commands in Terminal:

export GOOGLE_GEMINI_BASE_URL="https://api.4allapi.com"
export GEMINI_API_KEY="你的API密钥"
export GEMINI_MODEL="gemini-3-pro"

Permanent setup method

Add the following to your shell configuration file:

For zsh (default)

echo 'export GOOGLE_GEMINI_BASE_URL="https://api.4allapi.com"' >> ~/.zshrc
echo 'export GEMINI_API_KEY="你的API密钥"' >> ~/.zshrc
echo 'export GEMINI_MODEL="gemini-2.5-pro"' >> ~/.zshrc
source ~/.zshrc

For bash

echo 'export GOOGLE_GEMINI_BASE_URL="https://api.4allapi.com"' >> ~/.bash_profile
echo 'export GEMINI_API_KEY="你的API密钥"' >> ~/.bash_profile
echo 'export GEMINI_MODEL="gemini-2.5-pro"' >> ~/.bash_profile
source ~/.bash_profile

Verify Gemini CLI environment variables

Verify in Terminal:

echo $GOOGLE_GEMINI_BASE_URL
echo $GEMINI_API_KEY
echo $GEMINI_MODEL

3. Windows configuration

PowerShell setup method

Run the following commands in PowerShell:

$env:GOOGLE_GEMINI_BASE_URL = "https://api.4allapi.com"
$env:GEMINI_API_KEY = "你的API密钥"
$env:GEMINI_MODEL = "gemini-3-pro"

Permanent PowerShell setup (user-level)

Run the following commands in PowerShell:

# Set user-level environment variables (persistent)
[System.Environment]::SetEnvironmentVariable("GOOGLE_GEMINI_BASE_URL", "https://api.4allapi.com", [System.EnvironmentVariableTarget]::User)
[System.Environment]::SetEnvironmentVariable("GEMINI_API_KEY", "你的API密钥", [System.EnvironmentVariableTarget]::User)
[System.Environment]::SetEnvironmentVariable("GEMINI_MODEL", "gemini-2.5-pro", [System.EnvironmentVariableTarget]::User)

:bulb: After making these changes, you need to reopen the PowerShell window for them to take effect.

Verify Gemini CLI environment variables

Verify in PowerShell:

echo $env:GOOGLE_GEMINI_BASE_URL
echo $env:GEMINI_API_KEY
echo $env:GEMINI_MODEL

Core Command Reference

Basic interaction

gemini  # Enter interactive mode
gemini -p "解释这段代码" @./src/utils.js  # File analysis
gemini --media wireframe.png "转换为React组件"  # Multimodal processing

Session management

# View command help
/quit  # Exit and show session statistics
/chat save feature_dev  # Save current session
/chat resume feature_dev  # Resume session

Tool integration

!ls -la  # Run a system command
/tools  # Show available tools
gemini review --staged-files --format=checklist  # Git integration

Repository operations

gemini -p "@./ 总结项目架构"  # Analyze the entire directory
gemini -p "@src/components/Button.jsx 生成单元测试"  # Specific file
gemini "迁移Django项目到Python3.12"  # Break down complex tasks

Advanced Memory Management (/memory Command)

Command comparison

Command	Function	Use case
/memory add	Add a new instruction	gemini /memory add “Always reply in Chinese”
/memory show	Show current memory	Verify AI context
/memory refresh	Reload the configuration file	After updating GEMINI.md

Configuration file practice

1. Create GEMINI.md to define project rules:

# 项目规范
- 使用TypeScript严格模式
- 函数命名：camelCase
- API响应格式：{ data, error }

2. Refresh the configuration:

/memory refresh  # Reload all configuration
/memory show  # Verify loaded content

Supports automatic scanning of multi-level directories; enterprise projects can define unified team standards

State Recovery (/restore Command)

Workflow

# Enable checkpointing at startup
gemini --checkpointing

# After performing operations that may modify files
/restore  # View recoverable checkpoints

# Restore to a specific state
/restore checkpoint_42

Applicable scenarios:

1. Roll back after a failed code-generation experiment
2. Recover from file editing mistakes
3. Create safety snapshots before batch operations

Error Handling Guide

Common error solutions

Error type	Symptom	Solution
Authentication failure	API key not found	Check environment variables or reset config.json
Quota exceeded	429 Too Many Requests	Switch to another model or wait
Context overflow	No clear error code (performance degradation)	Reduce input size / process in chunks
File creation failed	Insufficient permissions	Run with administrator privileges or check directory permissions
Network error	Connection timed out	1. Verify the proxy 2. Use curl to test API reachability

Debugging tools

gemini -d  # Enable verbose debug mode
/stats  # View resource usage

---

4All API Footer

4All API · One-stop AI large model API aggregation platform | Pricing | Contact us

© 2025 4All API. All rights reserved.