N2N Memory
A specialized MCP server designed to eliminate memory pollution during AI-assisted cross-project development.
n2n-memory is a specialized Model Context Protocol (MCP) server designed to eliminate memory pollution during AI-assisted cross-project development. By persisting AI cognitive fragments directly within each project's directory, it transforms transient context into a permanent, version-controlled asset. This ensures that an AI assistant's understanding of a specific codebase stays local, secure, and easily shareable across development teams. Memory files are stored locally at [Project Root]/.mcp/memory.json, preventing context leakage between different projects. The server features a Git-friendly architecture that automatically sorts JSON keys to generate clean, readable diffs. It is tool-agnostic and works seamlessly with all MCP-enabled models including Claude, Gemini, GPT-4o/5, and DeepSeek V3. Built with a privacy-first design, it keeps all cognitive data local and isolated from third-party cloud storage.
Key Highlights
Git-Friendly Architecture
JSON data automatically sorted by key to generate clean, readable git diff outputs.
Universal Compatibility
Works with Claude 3.5/4, Gemini 1.5 Pro/Flash, GPT-4o/5, and DeepSeek V3.
Tool Agnostic
Uses standard .mcp naming convention, not tied to any specific AI brand or IDE plugin.
Assets for Your Code
Memory stays with source code; team members synchronize AI understanding by pulling the repo.
Privacy-First Design
All cognitive data stays local and isolated from third-party cloud storage.
Project-Level Isolation
Memory files stored at [Project Root]/.mcp/memory.json prevent context leakage between projects.
Quick Start
// Claude Desktop config
{
"mcpServers": {
"n2n-memory": {
"command": "npx",
"args": ["-y", "@datafrog-io/n2n-memory"]
}
}
}