qzhaojing/Trae_MEM

Trae Memory MCP Server

A Model Context Protocol (MCP) server that provides long-term memory capabilities for Trae AI (or any MCP client). It uses a local SQLite database with hybrid search (Keyword + Semantic) to store and retrieve code snippets, decisions, and context.

Features

• Long-term Memory: Store text content with optional tags.
• Hybrid Search: Combines Full-Text Search (SQLite FTS5) with Semantic Search (Sentence Transformers).
• Local & Private: All data is stored locally in memory.db.
• Fast: Uses a lightweight embedding model (all-MiniLM-L6-v2) and efficient vector operations.

Prerequisites

• Python 3.10 or higher
• Windows, Linux, or macOS

Installation

1. Clone the repository

git clone https://github.com/yourusername/trae_mem_mcp.git
cd trae_mem_mcp

2. Create a Virtual Environment (Recommended)

Windows:

python -m venv .venv
.venv\Scripts\activate

Linux/macOS:

python3 -m venv .venv
source .venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

Note: The first time you run the server, it will download the embedding model (approx. 90MB) from Hugging Face. If you are in a region with restricted access, set the HF_ENDPOINT environment variable to a mirror before running (handled automatically in the Windows batch script).

Configuration in Trae

1. Open Trae Settings -> MCP Servers.
2. Click "Add" (Edit Configuration).
3. Add the following configuration (replace G:\\... with the actual path to your start_server.bat):

{
  "mcpServers": {
    "Trae_Memory": {
      "command": "cmd.exe",
      "args": [
        "/c",
        "G:\\path\\to\\trae_mem_mcp\\trae_mem_mcp\\start_server.bat"
      ]
    }
  }
}

Make sure to use double backslashes \\ in the JSON path.

Usage

Once connected, you can ask Trae to:

• "Remember that we decided to use JWT for authentication."
• "Search memory for authentication details."
• "What did we save about the database schema?"

Project Structure

• trae_mem_mcp/: Source code package
  • server.py: MCP server entry point and tool definitions
  • db.py: SQLite database operations (FTS + Vector storage)
  • model.py: Embedding model wrapper (SentenceTransformers)
  • start_server.bat: Windows startup script
• memory.db: SQLite database (created on first run)

License

MIT