home / mcp / quickbooks mcp server

QuickBooks MCP Server

Queries QuickBooks data using natural language through Claude Desktop by running a local MCP server.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "nikhilgy-quickbooks-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "<absolute_path_to_quickbooks_mcp_folder>",
        "run",
        "main_quickbooks_mcp.py"
      ],
      "env": {
        "QUICKBOOKS_ENV": "'sandbox' or 'production'",
        "QUICKBOOKS_CLIENT_ID": "your_actual_client_id",
        "QUICKBOOKS_COMPANY_ID": "your_actual_company_id",
        "QUICKBOOKS_CLIENT_SECRET": "your_actual_client_secret",
        "QUICKBOOKS_REFRESH_TOKEN": "your_actual_refresh_token"
      }
    }
  }
}

This MCP server lets you query QuickBooks data using natural language inside Claude Desktop, giving you a secure, local-first way to access financial information through conversational prompts.

How to use

You will run the QuickBooks MCP Server locally and connect Claude Desktop as your MCP client. After setup, you can ask your assistant to retrieve data from QuickBooks such as accounts, bills, and customers. Use natural language queries like: Get all accounts from QuickBooks or Get all customers from QuickBooks. The server handles authentication with your QuickBooks credentials and serves data to Claude Desktop through the MCP channel.

When Claude Desktop is configured, every launch loads the latest QuickBooks API tools and documentation, ensuring your assistant uses up-to-date endpoints. If you encounter a delay on first start, give it a moment while essentials download in the background. You can perform common data queries such as listing accounts, bills, or customers directly from your AI assistant.

How to install

Prerequisites you need before installation:

Python 3.10 or higher
Access to Claude Desktop
Internet access for initial tool installation and API documentation

Step 1. Install uv on your operating system. Run the appropriate command for your platform.

# MacOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Step 2. Configure Claude Desktop to use the MCP server. Prepare your QuickBooks credentials in a local environment file as described, then configure Claude Desktop to point to the MCP runner.

{
  "mcpServers": {
    "QuickBooks": {
      "command": "uv",
      "args": [
        "--directory",
        "<absolute_path_to_quickbooks_mcp_folder>",
        "run",
        "main_quickbooks_mcp.py"
      ]
    }
  }
}

Launch and usage

After configuring Claude Desktop, relaunch it. The first start may take 10–20 seconds as the required packages and QuickBooks API tools download in the background. Each launch ensures the most recent QuickBooks API tools are available to your assistant.

Examples you can try with your assistant after setup include querying accounts, bills, and customers from QuickBooks.

Available tools

query_accounts

Query and return the list of accounts from QuickBooks.

query_bills

Query and return bills with optional date filtering.

query_customers

Query and return customer records from QuickBooks.