MCP Memos

🚀 MCP-Memos Memo Tool

MCP-Memos is an information recording and retrieval tool built on the MCP protocol. It is designed to provide users with the ability to quickly save and search for text information. It is especially suitable for developers to quickly record and search for information during their workflow (e.g., using the cursor), without the need to switch to other applications.

🚀 Quick Start

MCP-Memos is an information recording and retrieval tool that allows you to quickly save and search for text information. It's a great fit for developers who need to jot down and find information on the fly without switching apps.

✨ Features

🔍 Advanced Search with Powerful LLM Support

MCP-Memos leverages large language models for retrieval, offering the most powerful fuzzy search capabilities:

• Semantic Understanding: Retrieves content based on meaning rather than keywords.
• Context Awareness: Understands user needs even with incomplete descriptions.
• Natural Language Query: Allows searching like having a conversation, no special syntax required.
• Concept Matching: Retrieves information by understanding concepts rather than simple text matching.

Unlike traditional vector or text-based fuzzy searches, MCP-Memos fully utilizes the power of LLMs to truly understand the user's retrieval intent, making it the most effective information retrieval method currently available.

📚 Documentation

📦 Installation

Download the mcp-memos binary file suitable for your computer architecture from releases.

💻 Usage Examples

Configuration

Add MCP-Memos to the macPilotCli configuration file:

{
    "mcpServers": {
      "MCP-Memos":{
        "command": "path/to/mcp-memos",
        "env": {
          "LLM_TOKEN": "xxxxxx",
          "LLM_BASE_URL": "xxxxxxxx/v1",
          "ANTHROPIC_MODEL": "xxxxx"
        }
      }
    }
}

Basic Usage

Recording Information

When you need to save information during development, there's no need to switch to other applications:

Record this memo. It's {description}

{content}

Retrieving Information

Find previously saved information:

Find memos about {description}

⚠️ Important Note

The description can vary as MCP-Memos' LLM understands concepts rather than just keywords, so it can match similar content even with different descriptions.

LLM_BASE_URL is an optional parameter, with the default being https://api.anthropic.com.

ANTHROPIC_MODEL is also an optional parameter, with the default being claude-3-7-sonnet-20250219.

It will accelerate the implementation of the sampling feature, eliminating the need to configure all environment parameters.