FLOWS

Table of Contents

• Introduction
• Features
• Memory Architectures

Introduction

aiFlows focuses on creating flexible and scalable AI agents, based on the Actor Model. This repository adds memory systems, vector store frameworks, and tools to the library. The memory capability empowers to maintain personalized, context-aware interactions over time.

Features

• Memory Systems: Implement short-term and long-term memory for AI interactions.
• Vector Store Integrations: Supports Qdrant and Chroma for efficient memory retrieval.
• Tooling: Incorporates over 15 tools, including web scrapers and more.

Memory Architectures

Short-Term Memory

• Session-based persistence: Updates the context dynamically after every interaction.
• Fetch and Update Prompts: Efficient mechanisms to retrieve and store context in real-time.

Long-Term Memory

• Vector Store: Stores facts with metadata for embedding-based retrieval.
• Graph-Based Memory: Utilizes Neo4j to model entities and relationships with detailed metadata.