quickbooks-desktop-mcp
Provides safe, structured access to a local QuickBooks Desktop company file for reading and writing transactions, with automatic undo logging.
README
QBBridge
An MCP frontend for QuickBooks Desktop.
Gives any MCP-compatible client (Claude Desktop, Claude Code, or anything else that speaks MCP) safe, structured access to a local QuickBooks Desktop company file: read accounts/vendors/transactions, post new transactions, and modify or delete existing ones -- with every write automatically logged so it can be reverted.
QuickBooks Desktop's automation API (QBXML/QBFC) only works on the machine (or LAN) where QuickBooks is actually installed -- there is no cloud proxy for it. QBBridge runs locally, right next to QuickBooks, and your company file's data never has to leave that machine to be used by an MCP client.
Why this exists
QBXML is powerful but has real, undocumented sharp edges: several element orderings are order-sensitive in ways the schema alone doesn't tell you, and getting them wrong produces a bare COM parse exception with no indication of which field was the problem. This package exists so you don't have to rediscover those the hard way -- they're encoded and tested once, here.
What it gives you
- Read: chart of accounts, vendor list, a vendor's transaction history (so you can classify a new transaction against how this vendor has actually been booked before, not a guess), flexible transaction queries, and a duplicate/overlap checker.
- Write: add/modify/delete transactions (Check, Deposit, CreditCardCharge, CreditCardCredit) and vendors.
- Safety: every write is automatically recorded to an append-only undo log as it happens (not batched at the end -- a crash mid-run still leaves a usable log of everything that did complete). Any run can be reverted, in reverse order, with one call.
What it deliberately does NOT do
This server has no business-specific logic in it -- no classification rules, no confidence thresholds, no assumptions about what kind of business you run. Deciding which vendor maps to which account, when to trust a match versus flag it for a human, and how to parse a specific bank's statement format is policy, and policy differs by business. That belongs in the agent/skill calling this server, not in the server itself. This keeps the server itself universal.
The companion skill
The server is the mechanism; skills/qbbridge-bookkeeping/SKILL.md
is the judgment. It's a Claude skill that teaches an agent how to actually
use these tools well -- when to trust a vendor's transaction history vs.
flag it for a human, how to handle payment/credit lines vs. purchases,
and the safety habits (checking the open company file, checking for
duplicates) to follow before every write. Install the MCP server for the
plumbing; add this skill for the bookkeeping judgment on top of it.
Requirements
- Windows, with QuickBooks Desktop installed and the target company file open.
- Python 3.10+.
pywin32(installed automatically on Windows).
Install
pip install qbbridge-mcp
Configure as an MCP server
Point your MCP client at the qbbridge-mcp command. For Claude Desktop,
add to your MCP config:
{
"mcpServers": {
"qbbridge": {
"command": "qbbridge-mcp"
}
}
}
Safety notes
- Every write tool accepts an optional
expected_company_fileparameter. Pass the path you expect QuickBooks to have open (from a priorqb_company_infocall); the tool refuses to run if a different file is actually open. This is the single check that prevents posting to the wrong client's file. qb_delete_transactionrequires you to pass a full snapshot of the transaction (query it first) -- without it, a delete cannot be undone.- Undo logs are written to
./qbbridge_runsby default, or$QBBRIDGE_RUNS_DIRif set.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.