MARM-Systems

概述

MARM（Memory Accurate Response Mode）是一个通用的MCP服务器，支持跨平台AI内存、多代理协调和上下文共享。它提供语义搜索、持久化跨会话记忆和智能回忆功能，使AI代理能够一致地记住并基于先前工作进行构建。该服务器支持多种传输协议，包括HTTP、STDIO和WebSocket，使其与Claude、Gemini和Qwen等多种AI客户端兼容。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• search_memory

  Search through persistent memory using semantic similarity

  试试:Find information about the previous project requirements试试:What decisions were made about the API design?
• log_session

  Log current conversation session to persistent memory

  试试:Save this conversation for future reference试试:Document our discussion about the implementation
• create_notebook

  Create a new persistent notebook entry

  试试:Create a new notebook for project documentation试试:Start a tracking log for this issue
• query_memory

  Query structured memory with specific criteria

  试试:Find all code snippets related to authentication试试:Retrieve all project decisions from last week
• classify_content

  Automatically classify content into memory categories

  试试:Categorize this code snippet试试:Classify this conversation as project-related
• generate_summary

  Generate a semantic summary of logged conversations

  试试:Summarize our last discussion试试:Create a summary of all project decisions
• share_memory

  Share memory across different AI agents

  试试:Share this information with Claude试试:Make this decision available to other AI tools
• clear_memory

  Clear specific memories or reset entire memory pool

  试试:Remove old project data试试:Clear all temporary information
• export_memory

  Export memory data in various formats

  试试:Export my conversation history as JSON试试:Prepare a report of all project decisions
• auto_log

  Enable automatic logging of conversations

  试试:Turn on automatic session logging试试:Set up auto-logging for this project
• manage_sessions

  Manage persistent sessions across conversations

  试试:Show all active sessions试试:Resume the previous coding session
• configure_memory

  Configure memory storage and retrieval parameters

  试试:Set up memory for this specific project试试:Configure memory retention settings

可对比工具

安装

docker pull lyellr88/marm-mcp-server:latest
docker run -d --name marm-mcp-server -p 8001:8001 -v ~/.marm:/home/marm/.marm lyellr88/marm-mcp-server:latest
claude mcp add --transport http marm-memory http://localhost:8001/mcp

pip install marm-mcp-server==2.2.6
pip install -r marm-mcp-server/requirements.txt
python marm-mcp-server
claude mcp add --transport http marm-memory http://localhost:8001/mcp

pip install marm-mcp-server==2.2.6
pip install -r marm-mcp-server/requirements_stdio.txt
claude mcp add --transport stdio marm-memory-stdio python "路径/to/marm-mcp-server/server_stdio.py"
python marm-mcp-server/server_stdio.py

FAQ

MARM与其他内存解决方案有何不同？

MARM专为MCP服务器设计，在协议级别提供持久化内存。它可跨不同AI客户端工作，并维护无论使用哪种工具的会话连续性。

MARM如何处理身份验证？

MARM使用OAuth 2.0进行身份验证。开发环境中包含模拟凭据，生产环境可替换为适当的OAuth提供者。

MARM-Systems 对比