expense-tracker
Manages personal expenses through natural language with tools for adding, updating, searching, and summarizing expenses, backed by Supabase.
README
💰 Expense Tracker MCP Server
A production-ready Model Context Protocol (MCP) server for your personal expense tracker, built with FastMCP 3.x and backed by Supabase (PostgreSQL).
Connect it to Claude Desktop, Cursor, or any MCP-compatible client and manage your finances with natural language.
✨ Features
| Tool | What it does |
|---|---|
add_expense |
Record a new expense |
update_expense |
Modify any field of an existing expense |
delete_expense |
Permanently remove an expense |
get_expense |
Fetch a single expense by UUID |
list_expenses |
Filter expenses by date, category, shop, amount, payment method |
search_expenses |
Full-text search across title, shop, notes, category |
summarize_expenses |
Totals & % breakdown by category / month / shop |
get_monthly_report |
Full month report with per-category breakdown |
get_budget_status |
Compare spending vs a budget limit |
get_shops |
All shops you've visited + visit counts + total spent |
get_categories |
All unique categories in your data |
export_expenses |
Export to CSV |
Resources (live data snapshots):
expense://today— Today's spendingexpense://recent/{n}— Last N expensesexpense://categories— All categories in use
🚀 Quick Setup
1. Create the Supabase Table
Go to supabase.com → your project → SQL Editor and run:
CREATE TABLE expenses (
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
title TEXT NOT NULL,
amount NUMERIC(12, 2) NOT NULL CHECK (amount > 0),
category TEXT NOT NULL,
shop TEXT,
date DATE NOT NULL,
payment_method TEXT,
notes TEXT,
created_at TIMESTAMPTZ DEFAULT now(),
updated_at TIMESTAMPTZ DEFAULT now()
);
-- Auto-update updated_at on every row change
CREATE OR REPLACE FUNCTION update_updated_at()
RETURNS TRIGGER AS $$
BEGIN
NEW.updated_at = now();
RETURN NEW;
END;
$$ LANGUAGE plpgsql;
CREATE TRIGGER set_updated_at
BEFORE UPDATE ON expenses
FOR EACH ROW EXECUTE FUNCTION update_updated_at();
-- Indexes for fast filtering
CREATE INDEX idx_expenses_date ON expenses(date);
CREATE INDEX idx_expenses_category ON expenses(category);
CREATE INDEX idx_expenses_shop ON expenses(shop);
2. Get Your API Keys
In Supabase → Project Settings → API:
- Copy your Project URL
- Copy your anon / public key (or service role key for private use)
3. Create Your .env File
copy .env.example .env
Edit .env:
SUPABASE_URL=https://your-project-id.supabase.co
SUPABASE_KEY=your-anon-key-here
4. Install Dependencies
uv sync
5. Test the Server
# Start the MCP inspector (browser UI to test all tools)
uv run fastmcp dev main.py
🔌 Claude Desktop Integration
Add this to your claude_desktop_config.json
(usually at %APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"expense-tracker": {
"command": "uv",
"args": ["run", "python", "main.py"],
"cwd": "C:\\Users\\saina\\OneDrive\\Desktop\\expense-tracker-mcp-server"
}
}
}
Restart Claude Desktop. You'll see a 🔌 icon with expense-tracker tools available.
💬 Example Prompts
Once connected to Claude Desktop, try:
"Add an expense: ₹450 for lunch at Dominos today, paid by UPI"
"Show me all my Food expenses this month"
"How much did I spend in June 2026?"
"Am I over my ₹5000 Food budget this month?"
"Where do I spend money most often?"
"Search for all Zomato expenses"
"Give me a monthly report for June 2026"
"Export all expenses from this month as CSV"
"What's my total spending this week by category?"
🗂️ Project Structure
expense-tracker-mcp-server/
├── main.py # FastMCP server — all 12 tools + 3 resources
├── db.py # Supabase data access layer
├── models.py # Pydantic validation models
├── .env # Your secrets (never commit this!)
├── .env.example # Template for .env
└── pyproject.toml # Dependencies
🛡️ Security Notes
- Never commit your
.envfile (it's in.gitignore) - Use the anon key for personal use — it's safe for client-side access
- Use Row Level Security (RLS) in Supabase if you ever share the project
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。