web-rag-fusion-mcp-server

Multi-Provider RAG Server

🌍 中文文档 | English

A powerful RAG (Retrieval-Augmented Generation) server with dynamic embedding provider support via the Model Context Protocol (MCP). Switch between OpenAI and DashScope/Qwen providers at runtime without code changes.

✨ Key Features

• 🔄 Dynamic Provider Switching - Runtime switching between OpenAI and DashScope
• 🏗️ Multi-Index Support - Different providers for different document collections
• 🛡️ Robust Error Handling - Automatic fallback and comprehensive error recovery
• 🌐 Web Search Integration - Enhanced search via Perplexity/Exa APIs
• ⚙️ Environment-Based Config - Zero-code configuration changes
• 🚀 Production Ready - Rate limiting, metrics, and monitoring

🚀 Quick Start

1. Installation

git clone https://github.com/Renewdxin/multi-provider-rag.git
cd multi-provider-rag
pip install -r requirements.txt

2. Configuration

Copy and configure environment variables:

cp .env.example .env
# Edit .env with your API keys

Basic OpenAI Setup:

EMBED_PROVIDER=openai
OPENAI_API_KEY=your_openai_key_here
EMBEDDING_MODEL=text-embedding-3-small

DashScope/Qwen Setup:

EMBED_PROVIDER=dashscope
DASHSCOPE_API_KEY=your_dashscope_key_here
EMBEDDING_MODEL=text-embedding-v1
OPENAI_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1

3. Run the Server

python -m src.mcp_server

🔧 Usage Examples

Dynamic Provider Selection

from src.embedding_provider import get_embed_model

# Use OpenAI
openai_model = get_embed_model("openai")

# Use DashScope
dashscope_model = get_embed_model("dashscope", model="text-embedding-v1")

# Environment-based selection
embed_model = get_embed_model_from_env()  # Uses EMBED_PROVIDER

Multiple Indexes with Different Providers

from src.embedding_provider import create_index_with_provider

# Create specialized indexes
docs_index = create_index_with_provider("openai", documents)
code_index = create_index_with_provider("dashscope", code_docs)

🌐 Supported Providers

Provider	Models	Endpoint	Features
OpenAI	text-embedding-ada-002 text-embedding-3-small text-embedding-3-large	https://api.openai.com/v1	High quality, global availability
DashScope	text-embedding-v1 text-embedding-v2	https://dashscope.aliyuncs.com/compatible-mode/v1	China-optimized, cost-effective

⚙️ Configuration

Environment Variables

Variable	Description	Default	Required
EMBED_PROVIDER	Embedding provider (openai/dashscope)	openai	No
EMBEDDING_MODEL	Model name (provider-specific)	text-embedding-3-small	No
OPENAI_API_KEY	OpenAI API key	-	Yes (for OpenAI)
DASHSCOPE_API_KEY	DashScope API key	-	Yes (for DashScope)
SEARCH_API_KEY	Perplexity/Exa API key	-	Optional
VECTOR_STORE_PATH	Vector database path	./data/vector_store.db	No

Provider Switching

Switch providers instantly by updating environment variables:

# Switch to DashScope
export EMBED_PROVIDER=dashscope
export DASHSCOPE_API_KEY=your_key

# Switch to OpenAI  
export EMBED_PROVIDER=openai
export OPENAI_API_KEY=your_key

🐳 Docker Deployment

# Build and run
docker-compose up -d

# With custom configuration
docker-compose -f docker-compose.yml up -d

📖 API Reference

MCP Tools

• search_knowledge_base - Search local vector database
• web_search - Search web via Perplexity/Exa
• smart_search - Hybrid local + web search
• add_document - Add documents to knowledge base

Python API

# Core embedding functions
from src.embedding_provider import (
    get_embed_model,
    get_embed_model_from_env,
    create_index_with_provider,
    validate_provider_config
)

# RAG engine
from src.llamaindex_processor import RAGEngine

# MCP server
from src.mcp_server import RAGMCPServer

🔍 Provider Validation

Check your provider configuration:

from src.embedding_provider import validate_provider_config

# Validate OpenAI setup
openai_status = validate_provider_config("openai")
print(f"OpenAI ready: {openai_status['valid']}")

# Validate DashScope setup  
dashscope_status = validate_provider_config("dashscope")
print(f"DashScope ready: {dashscope_status['valid']}")

🚀 Production Features

• Rate Limiting - Configurable request throttling
• Monitoring - Prometheus metrics integration
• Logging - Structured logging with loguru
• Error Recovery - Automatic provider fallback
• Health Checks - Built-in validation endpoints

📊 Performance

• Provider Switching - Zero downtime switching
• Caching - Intelligent query engine caching
• Batch Processing - Optimized bulk operations
• Memory Efficient - Lazy loading and cleanup

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-feature
3. Commit changes: git commit -m 'Add amazing feature'
4. Push to branch: git push origin feature/amazing-feature
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔗 Links

• Documentation: Full documentation
• Releases: GitHub Releases
• Issues: Bug Reports & Feature Requests

⭐ Star History

If this project helps you, please consider giving it a star! ⭐

---

Built with ❤️ for the AI community