Skip to main content

​
Knowledge Base Infrastructure

The knowledge base infrastructure provides the foundation for agent memory, context management, and information retrieval across all MOOD MNKY agents.

​
Overview

The knowledge base system enables agents to:
  • Store and retrieve contextual information
  • Maintain conversation history and user preferences
  • Access structured and unstructured data
  • Support RAG (Retrieval-Augmented Generation) capabilities

​
Architecture

​
Vector Database

The knowledge base uses vector embeddings for semantic search:
  • Storage: Vector embeddings of documents and conversations
  • Search: Semantic similarity matching for context retrieval
  • Indexing: Automatic indexing of new content
  • Updates: Real-time synchronization across agents

​
Document Management

  • Ingestion: Automatic document processing and chunking
  • Metadata: Rich metadata tagging for enhanced search
  • Versioning: Document version control and history
  • Access Control: Permission-based access to knowledge

​
Integration Points

​
Configuration

Knowledge base configuration is managed through the agent infrastructure: 
knowledge_base:
  vector_store: "pinecone"
  embedding_model: "text-embedding-ada-002"
  chunk_size: 1000
  chunk_overlap: 200