expense-tracker
Enables tracking business trip expenses and calculating settlements using Google Sheets, with natural language parsing of expenses.
README
Expense Tracker MCP Server
A Model Context Protocol (MCP) server for tracking business trip expenses and calculating settlement amounts using Google Sheets.
Overview
This server provides tools to:
- Create Google Sheets for expense tracking
- Parse expenses from natural language input
- Track who paid what and calculate who owes whom
- Support multiple participants and settlement calculations
Prerequisites
- Node.js 18+
- Google Cloud project with Sheets API enabled
- Google OAuth 2.0 credentials (or Service Account)
Installation
npm install
npm run build
npm test # Verify all 18 tests pass
Setup
Option 1: Google OAuth 2.0 (Recommended for Small Projects)
- Create a Google Cloud project and enable Sheets + Drive APIs
- Create OAuth 2.0 credentials (Desktop application)
- Copy
.env.exampleto.envand add your credentials:GOOGLE_CLIENT_ID=your-client-id.apps.googleusercontent.com GOOGLE_CLIENT_SECRET=your-client-secret GOOGLE_REDIRECT_URI=http://localhost:3000/oauth/callback DEFAULT_PARTICIPANTS=Mike,Chris DEFAULT_CURRENCY=MXN - Complete OAuth authorization to get refresh token:
npm run build node get-refresh-token.js - Copy the generated
GOOGLE_REFRESH_TOKENto your.env
Option 2: Service Account (Better for Production)
- Create a Service Account in Google Cloud Console
- Download the JSON key file
- Update configuration to use Service Account credentials
Configuration
Create a .env file from .env.example with:
- Google OAuth credentials
- Default participants list
- Preferred currency
Tools
- create-expense-sheet - Create a new expense tracking spreadsheet
- add-expense - Add a single expense entry
- add-expenses-bulk - Parse and add multiple expenses
- get-expenses - Query expenses with filters
- get-settlement-summary - Calculate final settlement
Using with Claude
Add to your Claude Desktop config (~/.config/Claude/claude_desktop_config.json):
{
"mcpServers": {
"expense-tracker": {
"command": "node",
"args": ["/path/to/expense-tracker-mcp/dist/index.js"],
"env": {
"GOOGLE_CLIENT_ID": "your-client-id",
"GOOGLE_CLIENT_SECRET": "your-secret",
"GOOGLE_REDIRECT_URI": "http://localhost:3000/oauth/callback",
"GOOGLE_REFRESH_TOKEN": "your-refresh-token",
"DEFAULT_PARTICIPANTS": "Mike,Chris",
"DEFAULT_CURRENCY": "MXN"
}
}
}
}
Development
npm run watch # Watch TypeScript compilation
npm run dev # Build and start server
npm test # Run all tests (18 tests)
npm start # Start the server
Architecture
src/index.ts- MCP server entry point and tool handlerssrc/config.ts- Environment configuration with Zod validationsrc/types.ts- TypeScript type definitionssrc/google/- Google Sheets API integrationauth.ts- OAuth 2.0 authenticationclient.ts- Sheets API clienttypes.ts- Google API types
src/utils/- Utility functionsparsers.ts- Natural language expense parsingsettlement.ts- Settlement calculation logicformatters.ts- Output formatting
Security Notes
- Never commit
.envfile - it contains credentials - Never expose refresh tokens in code or logs
- Use environment variables for all sensitive data
- Consider Service Account for production deployments
Testing
All functionality is tested with 18 comprehensive tests:
npm test
License
MIT
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。