EMMS - Enhanced Memory Management System

What is EMMS - Enhanced Memory Management System?

EMMS is a cognitive memory system for AI agents, providing 129 MCP tools for storage, retrieval, knowledge graphs, reflection, and metacognition. It is available under the MIT License and can be installed via pip, npx, or uvx.

How to use EMMS - Enhanced Memory Management System?

Install the package using pip install emms-mcp, npx -y emms-mcp, or uvx emms-mcp. Configure it in an MCP client by adding a server entry with command uvx and args ["emms-mcp"]. The server exposes 129 tools across categories like Storage, Retrieval, Knowledge Graph, Reflection, Emotions, Goals, Metacognition, Memory Management, Identity, Prediction, and Multi-Agent.

Key features of EMMS - Enhanced Memory Management System

• Provides 129 MCP tools for cognitive memory tasks
• Supports hybrid and associative retrieval methods
• Includes knowledge graph building and analysis tools
• Offers reflection and dream-like synthesis capabilities
• Tracks emotional states and mood trends
• Enables multi-agent memory merging and migration

Use cases of EMMS - Enhanced Memory Management System

• AI agents maintaining long-term memory across sessions
• Building and querying knowledge graphs from stored information
• Performing metacognitive analysis and bias detection
• Coordinating shared memory across multiple AI agents
• Enabling adaptive retrieval based on context and affect

FAQ from EMMS - Enhanced Memory Management System

What platforms or runtimes does EMMS require?

It requires a Python environment (via pip) or Node.js (via npx), and can also be run via uvx. It must be used with an MCP-compatible client.

What types of memory operations does it support?

It supports storing, retrieving, decaying, deduplicating, and reconsolidating memories, along with advanced retrieval like hybrid, adaptive, and affective retrieval.

Can EMMS be used with multiple AI agents?

Yes, it includes multi-agent tools such as merge_from, list_namespaces, and migrate_namespace for sharing or moving memory between agents.

Does it support emotional modeling?

Yes, it provides tools for managing current emotion, regulating emotions, tracking mood trends, and viewing the emotional landscape.

Is there any built-in reflection or self-improvement?

Yes, the Reflection category includes reflect, dream, synthesize_wisdom, and abstract_principles tools, along with metacognitive reports and consciousness metrics.