@nmeierpolys
基本信息
Make it easy for agents to build their context about your projects over time
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"mcp-structured-memory": {
"command": "npx",
"args": [
"@nmeierpolys/mcp-structured-memory"
]
}
}
}Create a new structured memory document with optional initial content. IMPORTANT: After using this tool, you MUST show the user the complete installation instructions returned by the tool - the memory will not work without proper MCP server setup and project context configuration.
Add an item to a list section in a memory document
Retrieve a specific section from a memory document
List all available memory documents
Get a high-level summary of a memory document
Search for information within a memory document
Update an entire section of a memory document. Content supports full Markdown formatting including headings, **bold**, *italic*, `code blocks`, [links](url), lists, tables, and all standard Markdown syntax.
Update an existing item in a list section
Move an item from one section to another
Retrieve the complete content of a memory document with all Markdown formatting preserved (headings, **bold**, *italic*, `code`, [links](url), tables, lists, etc.)
Structured Memory is a Model Context Protocol (MCP) server that provides structured, domain-specific memory management through markdown files. It is designed for ongoing projects where you want to accumulate valuable context over time, such as travel planning, research projects, real estate search, investment theses, product planning, and career development.
Install via npm (npm install -g @nmeierpolys/mcp-structured-memory) or from source. Configure the server in your MCP client (e.g., Claude Desktop) by adding an entry to claude_desktop_config.json. Use tools like create_memory and list_memories to manage memory documents; the LLM will automatically update these documents as you converse.
Traditional MCP memory servers use semantic search across scattered conversation snippets. Structured Memory instead maintains living documents with structured content that you can scan, update, and track over time, like a personal notebook with AI assistance.
On macOS: `~/Library/Application Support/mcp-structured-memory/`. On Windows: `%LOCALAPPDATA%\mcp-structured-memory\`. On Linux: `~/.local/share/mcp-structured-memory/`.
Use the `create_memory` tool, or ask your LLM client to create one, for example: “Create a new travel advisor memory document and tell me how to use it.”
Yes, the server automatically creates timestamped backups before every major update to a memory document.
The server provides ten tools: `create_memory`, `list_memories`, `get_memory_summary`, `get_section`, `get_full_memory`, `search_within_memory`, `update_section`, `add_to_list`, `update_list_item`, and `move_list_item`.
Local-first MCP server that snapshots your project's working state into one context object. Open-core, privacy-first.
The ultimate RAG for your monorepo. Query, understand, and edit multi-language codebases with the power of AI and knowledge graphs
One permissioned folder your team and your AI agents both log into.
Orcha is a unified context layer for AI tools and agents. It stores and indexes and exposes organizational, team, or individual knowledge. Files, structured databases, and connected sources live in one workspace, and age
The Cloud is the AI-native workspace that gives every AI — Claude, ChatGPT, Gemini, and other Frontier & Free or Local models — shared memory and real agents. Not cloud computing: thecloud.so is one workspace where your
评论