Math I Choose to Ignore MCP Server
AI-powered expense tracking MCP server that connects LLM clients to PostgreSQL for natural language expense management.
README
Math I Choose to Ignore (MCP Server)
Because your bank account won't ignore it forever.
Math I Choose to Ignore is an AI-powered expense tracking server built using the Model Context Protocol (MCP). It connects LLM clients (like Claude Desktop) directly to your PostgreSQL database, transforming your AI assistant into a personal financial analyst that can track, categorize, summarize, and manage your spending through natural language.
Features
- Natural Language Ledger: Add, edit, delete, and query your expenses just by chatting with Claude.
- Robust PostgreSQL Integration: Uses strict financial schema design (
NUMERIC(10,2)for currency precision, nativeDATEtypes, andSERIALprimary keys). - Safe Dynamic Querying: Built-in SQL injection prevention using parameterized queries and dynamic condition building.
- Built with FastMCP: Exposes both MCP Tools (for database actions) and MCP Resources (for read-only context like expense categories).
- Isolated Environment: Clean dependency management using
uv.
Tech Stack
- Language: Python
- MCP Framework:
fastmcp - Database: PostgreSQL (
psycopg/psycopg2) - Package Manager:
uv
Getting Started
1. Prerequisites
- Python 3.10+
- uv package manager
- PostgreSQL installed and running
- Claude Desktop (for client integration)
2. Installation & Setup
Clone the repository and install dependencies:
git clone [https://github.com/yourusername/math-i-choose-to-ignore.git](https://github.com/yourusername/math-i-choose-to-ignore.git)
cd math-i-choose-to-ignore
uv venv
uv sync
Database Configuration
DB_URL = "dbname=yourdbname user=yourusername password=yourpassword host=localhost port=5432"
Testing with MCP Inspector
uv run fastmcp dev inspector main.py
Claude Desktop Integration
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"Math I Choose to Ignore": {
"command": "C:\\Users\\YourUsername\\.local\\bin\\uv.exe",
"args": [
"--directory",
"C:\\Users\\YourUsername\\Desktop\\MCP\\math_i_choose_to_ignore",
"run",
"--with",
"fastmcp",
"fastmcp",
"run",
"C:\\Users\\YourUsername\\Desktop\\MCP\\math_i_choose_to_ignore\\main.py"
]
}
}
}
Important: After updating the JSON file, fully quit Claude Desktop from your Windows system tray and restart it.
Available MCP Tools And Resources
Tools (Actions)
| Tool | Description |
|---|---|
add_expense |
Adds a new expense with date, amount, category, subcategory, and optional notes. |
list_exp |
Lists all expenses within an inclusive YYYY-MM-DD date range. |
summarize |
Groups and sums expenses by category over a date range (optional category filter). |
edit_exp |
Updates an existing expense amount or record. |
delete_exp |
Safely removes an expense entry from the ledger. |
Resources (Read-Only Context)
| Resource URI | Description |
|---|---|
expense://categories |
Exposes standard budget categories from a local JSON file so the AI knows valid categorization options. |
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。