RAG & Document Loading

Using document loaders and vector memory with agents to provide grounded, factual responses from your knowledge base.

Agents can leverage document loaders and vector memory to access knowledge bases and provide grounded, factual responses.

🔄 Agent RAG Workflow

Basic RAG Agent

// Step 1: Create vector memory
vectorMemory = aiMemory( "chroma", {
    collection       : "product_docs",
    embeddingProvider: "openai",
    embeddingModel   : "text-embedding-3-small"
} )

// Step 2: Ingest documents
result = aiDocuments( "/docs/products", {
    type      : "directory",
    recursive : true,
    extensions: [ "md", "txt", "pdf" ]
} ).toMemory(
    memory  : vectorMemory,
    options : { chunkSize: 1000, overlap: 200 }
)

println( "Loaded #result.documentsIn# documents as #result.chunksOut# chunks" )

// Step 3: Create agent with vector memory
agent = aiAgent(
    name        : "Product Support",
    description : "Product documentation specialist",
    instructions: "Answer questions using the product documentation. Always cite sources.",
    memory      : vectorMemory
)

// Step 4: Query — agent automatically retrieves relevant docs
response = agent.run( "How do I configure SSL certificates?" )

Multi-Source RAG Agent

Combine multiple knowledge bases by passing an array of memories:

productDocs = aiMemory( "chroma", { collection: "product_docs" } )
apiDocs     = aiMemory( "chroma", { collection: "api_docs" } )
faqMemory   = aiMemory( "chroma", { collection: "faq" } )

// Ingest each source
aiDocuments( "/docs/products", { type: "directory" } ).toMemory( productDocs )
aiDocuments( "/docs/api",      { type: "directory" } ).toMemory( apiDocs )
aiDocuments( "/docs/faq.md",   { type: "markdown"  } ).toMemory( faqMemory )

// Agent searches across all knowledge bases
agent = aiAgent(
    name    : "Knowledge Assistant",
    memories: [ productDocs, apiDocs, faqMemory ]
)

response = agent.run( "Explain the authentication API" )

RAG Agent with Real-Time Tools

Combine document retrieval with live data access:

docMemory = aiMemory( "chroma", { collection: "documentation" } )
aiDocuments( "/docs", { type: "directory" } ).toMemory( docMemory )

statusTool = aiTool(
    "check_system_status",
    "Check current system status and metrics",
    () => getCurrentSystemStatus()
)

agent = aiAgent(
    name        : "System Assistant",
    instructions: "Use docs for general questions. Use the status tool for real-time data.",
    memory      : docMemory,
    tools       : [ statusTool ]
)

// Agent selects the right approach per query
agent.run( "What are the system requirements?" )  // → uses docs
agent.run( "Is the system currently running?" )   // → uses statusTool

📚 For Full RAG Capabilities

This page covers agent-level RAG setup. For deeper detail on:

• Document loaders (PDF, CSV, JSON, XML, HTTP, web crawlers, SQL, etc.) — see Document Loaders

• Chunking strategies, batch loading, async ingestion — see RAG Guide

• Vector memory configuration (Pinecone, Chroma, Qdrant, pgvector, etc.) — see Vector Memory Systems

Related Pages

• Memory Management — Memory types and per-call identity routing

• Vector Memory Systems — Full vector store configuration

• Document Loaders — All loader types

• RAG Guide — End-to-end RAG implementation