home / mcp / cursor db mcp server

Cursor DB MCP Server

Model Context Protocol server for querying Cursor chat history

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "jbdamask-cursor-db-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "<fully qualified path to>/cursor-db-mcp-server.py"
      ],
      "env": {
        "YOUR_ENV_PLACEHOLDER": "VALUE"
      }
    }
  }
}

You can access Cursor IDE’s SQLite data through a dedicated MCP server that exposes project data, chat history, and composer information to AI assistants. This enables you to query and analyze Cursor data from compatible MCP clients, making it easier to explore your work, conversations, and composer's data in a structured way.

How to use

You will use an MCP client to connect to Cursor DB MCP Server. Once connected, you can request lists of Cursor projects, retrieve chat data for a specific project, and get composer IDs for projects. You can also query specific tables within a project’s database and refresh the list of database paths when new data is added.

How to install

Prerequisites: you need Python installed on your system and Cursor IDE set up.

Step 1: Install the MCP server dependencies and setup using the provided installer script.

Additional sections

Configuration, security, and practical usage notes follow to help you integrate Cursor DB MCP Server smoothly into your workflow.

Configuration examples

Below are explicit MCP configurations you can use with MCP clients. They show local (stdio) connections using uv as the runner. Use the exact commands and paths shown.

Security and notes

Keep MCP endpoints secure and restrict access to trusted AI assistants. The server reads Cursor data from your installation directories and exposes it through MCP resources and tools. If you store sensitive project data, consider network access controls and credential management.