🍙 A personal AI agent & local memory hub for all AI agents like Claude Code, Codex, OpenClaw and Her

This project presents a personal AI agent framework that serves as a centralized local memory hub for multiple AI agents (Claude, Codex, OpenClaw, Hermes). It enables different AI systems to share context, maintain persistent memory, and coordinate interactions through a unified interface. The system acts as a middleware layer that enhances individual AI agents with shared knowledge and state management capabilities.

Key Points

• •Create a unified memory hub that multiple AI agents can access and contribute to simultaneously
• •Implement persistent local storage for conversation history, user preferences, and agent state
• •Enable context sharing between different AI models (Claude, Codex, OpenClaw, Hermes) to avoid redundant explanations
• •Design a middleware architecture that intercepts and routes agent communications through the memory system
• •Establish standardized data formats for memory entries to ensure compatibility across different AI agents
• •Implement conflict resolution when multiple agents attempt to update the same memory segments
• •Create agent-specific memory partitions while maintaining a shared global context layer
• •Build a query interface for agents to retrieve relevant historical context before responding to users

Workflow Diagram

Concepts