@hyflog/gemini-cli

Enhanced Gemini CLI with multi-provider support, tool calling, and advanced automation features.

🚀 Installation

# Install globally
npm install -g @hyflog/gemini-cli

# Or use with npx (no installation required)
npx @hyflog/gemini-cli --help

✨ Features

• Multi-Provider Support: OpenAI, DeepSeek, Gemini, and custom endpoints
• Tool Calling: Automatic command execution with --yolo flag
• File Integration: Include files with @filename syntax
• Shell Commands: Execute commands with !command syntax
• Slash Commands: Built-in commands like /help, /tools, /stats
• Streaming Responses: Real-time output for better UX
• Custom Base URLs: Support for proxy services and self-hosted models

🔧 Quick Start

1. Basic Usage

# Simple text conversation
echo "Hello, introduce yourself" | gemini --provider openai-compatible --model-name gpt-4o

# With API configuration
OPENAI_API_KEY=your-key \
OPENAI_BASE_URL=https://api.openai.com/v1 \
echo "Hello" | gemini --provider openai-compatible --model-name gpt-4o

2. Tool Calling (Automation)

# Manual confirmation (safe mode)
echo "Create a test folder on desktop" | gemini --provider deepseek --model-name deepseek-chat

# Automatic execution (--yolo flag)
echo "Create a test folder on desktop" | gemini --provider deepseek --model-name deepseek-chat --yolo

3. File Integration

# Include file content in conversation
echo "@package.json Analyze this project" | gemini --provider openai-compatible --model-name gpt-4o

4. Shell Commands

# Execute shell commands
echo "!ls -la" | gemini --provider openai-compatible --model-name gpt-4o

🌐 Supported Providers

OpenAI Compatible

# OpenAI Official
OPENAI_API_KEY=sk-your-key \
OPENAI_BASE_URL=https://api.openai.com/v1 \
gemini --provider openai-compatible --model-name gpt-4o

# Third-party proxy
OPENAI_API_KEY=sk-proxy-key \
OPENAI_BASE_URL=https://your-proxy.com/v1 \
gemini --provider openai-compatible --model-name gpt-4o

DeepSeek

DEEPSEEK_API_KEY=sk-your-deepseek-key \
gemini --provider deepseek --model-name deepseek-chat

Self-hosted Models

# Local Ollama
OPENAI_BASE_URL=http://localhost:11434/v1 \
gemini --provider openai-compatible --model-name llama3.2

# Custom server
OPENAI_BASE_URL=http://your-server:port/v1 \
gemini --provider openai-compatible --model-name "your-model"

📖 Command Reference

Core Options

• --provider - Model provider (openai-compatible, deepseek, gemini)
• --model-name - Model name (gpt-4o, deepseek-chat, etc.)
• --api-key - API key (or use environment variables)
• --base-url - API base URL for OpenAI-compatible services
• --yolo - Enable automatic tool execution
• --telemetry false - Disable telemetry

Built-in Commands

• /help - Show help information
• /tools - List available tools
• /stats - Show session statistics
• /clear - Clear conversation history
• @filename - Include file content
• !command - Execute shell command

🔑 Environment Variables

# OpenAI Compatible
export OPENAI_API_KEY=your-api-key
export OPENAI_BASE_URL=https://api.openai.com/v1

# DeepSeek
export DEEPSEEK_API_KEY=your-deepseek-key

# Gemini (Google)
export GOOGLE_API_KEY=your-google-key

📚 Examples

Development Workflow

# Project analysis
echo "@package.json @README.md Analyze this project structure" | \
gemini --provider openai-compatible --model-name gpt-4o --yolo

# Code review
echo "!git diff HEAD~1" | \
gemini --provider deepseek --model-name deepseek-coder

# Automated tasks
echo "Create a new React component called UserProfile" | \
gemini --provider openai-compatible --model-name gpt-4o --yolo

System Administration

# System monitoring
echo "Check system performance and suggest optimizations" | \
gemini --provider deepseek --model-name deepseek-chat --yolo

# Log analysis
echo "@/var/log/nginx/error.log Analyze these errors" | \
gemini --provider openai-compatible --model-name gpt-4o

🛠️ Advanced Configuration

Custom Script

Create a helper script for easier usage:

#!/bin/bash
# gemini-helper.sh

case $1 in
  "openai")
    echo "$2" | OPENAI_API_KEY=$OPENAI_API_KEY \
    OPENAI_BASE_URL=https://api.openai.com/v1 \
    gemini --provider openai-compatible --model-name gpt-4o --telemetry false
    ;;
  "deepseek")
    echo "$2" | DEEPSEEK_API_KEY=$DEEPSEEK_API_KEY \
    gemini --provider deepseek --model-name deepseek-chat --telemetry false
    ;;
  *)
    echo "Usage: $0 {openai|deepseek} 'your question'"
    ;;
esac

🐛 Troubleshooting

Common Issues

1. Permission Denied

   chmod +x $(which gemini)

2. API Key Errors

   # Check environment variables
   echo $OPENAI_API_KEY
   echo $OPENAI_BASE_URL

3. Network Issues

   # Test API connectivity
   curl -H "Authorization: Bearer $OPENAI_API_KEY" $OPENAI_BASE_URL/models

📄 License

MIT License - see LICENSE file for details.

🤝 Contributing

Issues and pull requests are welcome at: https://github.com/hyflog/gemini-cli

📞 Support

• 📖 Documentation: Provider Usage Guide
• 🐛 Issues: GitHub Issues
• 💬 Discussions: GitHub Discussions