Naive RAG

Naive RAG uses pure vector (embedding) similarity to find relevant document chunks. It's the fastest architecture to deploy and works well for simple Q&A over a single domain.

How it Works

1. Your query is converted to an embedding vector
2. The vector database finds the most similar chunk embeddings
3. Top-K results are returned ranked by similarity score

RAG Wizard Configuration

Step 1: Project Setup

Configure your project with a focused domain:

• Domain: Select one specific domain (e.g., "Customer Support")
• Use Case: Choose "Document Analysis" or "Q&A"

Step 2: Data Sources

Upload documents from a single domain for best results. Mix multiple topics and the similarity scores will drop.

Step 3: Document Processing

Step 4: Pipeline Configuration

Step 5: API Endpoints

Generate your API key and integrate using REST API, Python SDK, or MCP Protocol:

curl -X POST "https://api.guidedmind.ai/rag/search" \
-H "X-API-Key: rk_your_key_here" \
-H "Content-Type: application/json" \
-d '{
  "query": "What is the refund policy?",
  "options": {
    "limit": 5,
    "threshold": 0.7
  }
}'

When to Use Naive RAG

When to Upgrade

Consider switching to Hybrid RAG if:

• Users search with exact keywords (product codes, error messages)
• Similarity scores consistently stay below 0.5
• You need to cover both semantic and literal matches

Consider Graph RAG if:

• Your documents reference each other
• Users ask multi-hop questions ("How does X relate to Y?")
• You need to discover hidden relationships