Configuration

Detailed configuration options and best practices for SmartRAG

Configuration Guide

Configure SmartRAG for your specific needs with detailed options and best practices

Basic Configuration

SmartRAG can be configured with various options to suit your needs:

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.Anthropic;
    options.StorageProvider = StorageProvider.Qdrant;
    options.ApiKey = "your-api-key";
});

Configuration Options

Option	Type	Default	Description
AIProvider	AIProvider	Anthropic	The AI provider to use for embeddings
StorageProvider	StorageProvider	Qdrant	The storage provider for vectors
ApiKey	string	Required	Your API key for the AI provider
ModelName	string	Provider default	The specific model to use
ChunkSize	int	1000	Size of document chunks
ChunkOverlap	int	200	Overlap between chunks

AI Provider Configuration

Choose from multiple AI providers for embedding generation:

Anthropic

OpenAI

Azure OpenAI

Gemini

Custom

Anthropic (Claude)

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.Anthropic;
    options.ApiKey = "your-anthropic-key";
    options.ModelName = "claude-3-sonnet-20240229";
});

OpenAI

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.OpenAI;
    options.ApiKey = "your-openai-key";
    options.ModelName = "text-embedding-ada-002";
});

Azure OpenAI

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.AzureOpenAI;
    options.ApiKey = "your-azure-key";
    options.Endpoint = "https://your-resource.openai.azure.com/";
    options.ModelName = "text-embedding-ada-002";
});

Google Gemini

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.Gemini;
    options.ApiKey = "your-gemini-key";
    options.ModelName = "embedding-001";
});

Custom AI Provider

services.AddSmartRAG(options =>
{
    options.AIProvider = AIProvider.Custom;
    options.CustomEndpoint = "https://your-custom-api.com/v1/embeddings";
    options.ApiKey = "your-custom-key";
    options.ModelName = "your-custom-model";
});

Storage Provider Configuration

Choose the storage backend that best fits your needs:

Qdrant

Redis

SQLite

In-Memory

File System

Qdrant (Vector Database)

services.AddSmartRAG(options =>
{
    options.StorageProvider = StorageProvider.Qdrant;
    options.QdrantUrl = "http://localhost:6333";
    options.CollectionName = "smartrag_documents";
});

Redis (In-Memory Cache)

services.AddSmartRAG(options =>
{
    options.StorageProvider = StorageProvider.Redis;
    options.RedisConnectionString = "localhost:6379";
    options.DatabaseId = 0;
});

SQLite (Local Database)

services.AddSmartRAG(options =>
{
    options.StorageProvider = StorageProvider.Sqlite;
    options.ConnectionString = "Data Source=smartrag.db";
});

In-Memory (Development)

services.AddSmartRAG(options =>
{
    options.StorageProvider = StorageProvider.InMemory;
    // No additional configuration needed
});

File System (Local Storage)

services.AddSmartRAG(options =>
{
    options.StorageProvider = StorageProvider.FileSystem;
    options.StoragePath = "./data/smartrag";
});

Advanced Configuration

Fine-tune SmartRAG for your specific requirements:

Custom Chunking

services.AddSmartRAG(options =>
{
    options.ChunkSize = 500;
    options.ChunkOverlap = 100;
    options.ChunkingStrategy = ChunkingStrategy.Sentence;
});

Document Processing

services.AddSmartRAG(options =>
{
    options.SupportedFormats = new[] { ".pdf", ".docx", ".txt" };
    options.MaxFileSize = 10 * 1024 * 1024; // 10MB
    options.EnableTextExtraction = true;
});

Environment Configuration

Configure SmartRAG using environment variables or configuration files:

appsettings.json

{
  "SmartRAG": {
    "AIProvider": "Anthropic",
    "StorageProvider": "Qdrant",
    "ApiKey": "your-api-key",
    "ChunkSize": 1000,
    "ChunkOverlap": 200
  }
}

Environment Variables

export SMARTRAG_AI_PROVIDER=Anthropic
export SMARTRAG_STORAGE_PROVIDER=Qdrant
export SMARTRAG_API_KEY=your-api-key

Best Practices

API Keys

Never hardcode API keys in source code. Use environment variables or secure configuration.

Chunk Size

Balance between context and performance. Smaller chunks for precision, larger for context.

Storage

Choose storage provider based on your scale and requirements.

Security

Use appropriate access controls and monitoring for production environments.