Dual memory system with expansive relevance search

Dual Memory System with Expansive Relevance Search

This repository provides a robust implementation of a dual memory system designed for AI developers. Leveraging PostgreSQL with the pgvector extension and Gemini Embeddings, it enables sophisticated relevance search capabilities. This tool is ideal for building AI agents and applications that require efficient and accurate retrieval of information from large datasets.

What it Does

The dual memory system architecture separates short-term context from long-term knowledge storage. Short-term memory handles immediate conversational context or task-specific data, while long-term memory stores a broader knowledge base. The system utilizes Gemini Embeddings to convert text into vector representations, which are then stored and indexed in PostgreSQL using pgvector. This allows for highly efficient and semantically relevant searches across the entire knowledge base, going beyond simple keyword matching.

Key Features

• Dual Memory Architecture: Distinct separation of short-term and long-term memory for optimized data management.
• PostgreSQL Backend: Utilizes a reliable and scalable relational database for data persistence.
• pgvector Integration: Enables efficient storage and querying of vector embeddings for advanced similarity search.
• Gemini Embeddings: Leverages powerful AI models for high-quality text-to-vector conversion, ensuring semantic understanding.
• Expansive Relevance Search: Facilitates finding information based on conceptual similarity rather than exact keyword matches.
• Open Source (GitHub): Full access to the source code for customization and integration.

Who it's For

This tool is specifically designed for AI developers, researchers, and engineers working on projects that require advanced memory management and information retrieval. It is particularly useful for:

• AI Agent Development: Building agents that can recall past interactions and access a broad knowledge base.
• RAG (Retrieval-Augmented Generation) Systems: Enhancing LLM responses with relevant information retrieved from external data.
• Knowledge Management Systems: Creating applications that need to organize and search large volumes of unstructured data.
• Personalized AI Experiences: Developing systems that adapt and learn from user interactions over time.