Add Support for any AI Provider ( Ant & Gem current ) #2
Description
🎯 Overview
This PR adds support for Google Gemini as an additional AI provider alongside the existing Anthropic Claude integration, giving users more choice in AI models for their MCP applications.
✨ Features Added
- Multi-AI Provider Support
- Google Gemini Integration: Full support for Gemini 1.5 Pro model
- Provider Selection: Easy switching between Anthropic Claude and Google Gemini
- Backward Compatibility: Existing Claude-only setups continue to work unchanged
- New Files
- python-example/client/client_multi_ai.py - Enhanced client with dual AI provider support
- python-example/client/run_client.py - Launcher script for easy provider selection
- python-example/client/SETUP_SUMMARY.md - Comprehensive setup documentation
- CHANGELOG.md - Project changelog
- Enhanced Configuration
- Updated .env file to support both API keys
- Added google-generativeai dependency to pyproject.toml
- Command-line arguments for provider selection
**🚀 Usage Examples**Quick Start
## Run with Claude (default)
- python run_client.py --provider anthropic
## Run with Gemini - python run_client.py --provider gemini
## List available providers - python run_client.py --list-providers
- Environment Configuration
## Choose your AI provider by setting one of these API keys - ANTHROPIC_API_KEY=your_anthropic_api_key
- GEMINI_API_KEY=your_gemini_api_key
## Set the provider you want to use: 'anthropic' or 'gemini' - AI_PROVIDER=anthropic
🔧 Technical _Implementation
_Gemini Tool Calling
- Uses prompt engineering approach for tool invocation (Gemini doesn't natively support MCP tools)_
- JSON-based tool calling interface
- Comprehensive error handling and fallbacks
- Architecture
- Maintains existing client architecture
- Polymorphic AI provider interface
- Clean separation of concerns between providers
- Compatibility
- Python 3.11+ (reduced from original 3.12+ requirement)
- All existing functionality preserved
- No breaking changes to existing APIs
🧪 Testing
Weather Queries Tested
✅ "What's the weather like in New York City?"
✅ "Are there any weather alerts in California?"
✅ "Tell me the forecast for Boston tomorrow."
Provider Switching
✅ Dynamic provider selection via command line
✅ Environment variable configuration
✅ Proper error handling for missing API keys
📚 Documentation Updates
Updated README.md with dual provider instructions
Added comprehensive setup documentation
Created changelog for tracking changes
Enhanced .gitignore for better project hygiene
🎯 Benefits
User Choice: Developers can choose their preferred AI provider
Cost Optimization: Switch between providers based on pricing/features
Redundancy: Fallback options if one provider has issues
Innovation: Easy to add more providers in the future
🔍 Code Review Notes
All new code follows existing patterns and conventions
Comprehensive error handling and logging
Type hints and documentation maintained
No impact on existing functionality
_
***This enhancement makes the MCP Streamable HTTP example more versatile and demonstrates how MCP can work with multiple AI providers, not just Anthropic Claude.
_