What is 'Memory' for LLMs?

'Memory' is an MCP (Model Context Protocol) server designed to provide persistent storage and management for conversational memories of Large Language Models (LLMs).

How does Memory help LLMs maintain context?

It allows LLMs to save, retrieve, append, and clear their conversational history, ensuring they maintain crucial context across interactions and sessions for more coherent dialogue.

What database does Memory use for storage?

Memory utilizes MongoDB for robust, persistent storage of all LLM conversational data, including timestamps, LLM identifiers, and optional user IDs.

Can I update memories without overwriting everything?

Yes, Memory offers an 'add-memories' tool specifically designed to append new conversational data to existing stored memories without overwriting the entire history.

Memory

Name: Memory
Author: JamesANZ

•

Manages the storage and retrieval of conversational memories for multiple large language models.

Memory is a Model Context Protocol (MCP) server designed to persistently log and retrieve conversational memories from various Large Language Models (LLMs). It offers a robust solution for enhancing LLM capabilities by allowing them to store, retrieve, append, and clear user-specific conversation data. Utilizing MongoDB for storage, it ensures data persistence and provides a structured way to manage LLM interactions, enabling richer and more context-aware conversations.

主要功能

010 GitHub stars

02Retrieve all stored memories

03Append new memories without overwriting

04Clear all stored memories

05Persistent storage using MongoDB

06Save and overwrite LLM memories

使用案例

01Storing ongoing conversation context for an LLM to refer back to

02Allowing LLMs to retrieve historical interactions for personalized responses

03Enabling LLMs to add new data points to existing memory sets