Expense Tracker MCP Server Tooling Backend For LLM Agents

🧾 Expense Tracker MCP Server (Tooling Backend for LLM Agents)

A lightweight Expense Tracker + Budget Manager built as an MCP (Model Context Protocol) server using FastMCP, enabling LLM/AI agents to manage expenses, analyze spending trends, and monitor budgets through tool calling.

This project exposes expense and budget operations as MCP tools, backed by a SQLite database for persistence.

🚀 Features ✅ Expense Management

Add new expenses (amount, category, subcategory, note)

Update existing expenses

Delete expenses

List expenses within a date range

📊 Analytics & Insights

Summarize spending by category

Expense statistics:

total transactions

total spend

average spend

Spending trend analysis grouped by:

day

week

month

year

💰 Budgeting Tools

Set category budgets for a date range

Check budget usage:

amount spent

remaining budget

percentage used

over-budget detection

📤 Export

Export expenses in a date range to CSV file

📌 Resources

Exposes categories via MCP resource endpoint:

expense://categories

🏗️ Project Architecture

FastMCP server exposes tools callable by AI agents

SQLite3 database stores expenses and budgets

Tools act like an API layer for:

CRUD operations

analytics

reporting

budget monitoring

📂 Database Schema expenses table Column Type Description id INTEGER Primary key amount REAL Expense amount category TEXT Expense category date TEXT Date (YYYY-MM-DD) note TEXT Optional note subcategory TEXT Optional subcategory budgets table Column Type Description id INTEGER Primary key category TEXT Budget category amount REAL Budget amount start_date TEXT Start date end_date TEXT End date 🛠️ Tools Available (MCP) Expense Tools

add_expense(date, amount, category, subcategory='', note='')

list_expenses(start_date, end_date)

update_expense(id, date=None, amount=None, category=None, subcategory=None, note=None)

delete_expense(id)

Analytics Tools

summarize(start_date, end_date)

get_statistics(start_date, end_date, category=None)

get_spending_trends(start_date, end_date, group_by='month', category=None)

Budget Tools

set_budget(category, amount, start_date, end_date)

check_budget(category, start_date, end_date)

Export

export_expenses_csv(start_date, end_date, file_path)

MCP Resource

expense://categories → loads categories from categories.json

▶️ Running the MCP Server 1️⃣ Clone the repo git clone https://github.com/Ssoumay11/Expense-Tracker-MCP-Server-Tooling-Backend-for-LLM-Agents.git cd expense-tracker-mcp

2️⃣ Install dependencies pip install -r requirements.txt

3️⃣ Start the MCP server python app.py

(Replace app.py with the filename where your mcp.run() lives.)

🧪 Example Usage Add an expense { "date": "2026-01-11", "amount": 150.0, "category": "Food", "subcategory": "Dining Out", "note": "Lunch with friends" }

Summarize spending

Request:

{ "start_date": "2026-01-01", "end_date": "2026-01-31" }

Output:

[ {"category": "Food", "total_amount": 2500.0}, {"category": "Housing", "total_amount": 12000.0} ]

🔒 Notes

Dates should be passed in the format: YYYY-MM-DD

SQLite DB is stored locally as: data.db

Categories are loaded from: categories.json

🧠 Why This Project Matters (AI Engineer Angle)

This project demonstrates:

Building a tool backend for AI agents via MCP

Designing reliable, structured tool interfaces

Persistent storage + analytics for agent decision-making

Real-world agent automation workflows (tracking, reporting, budgeting)

🛠️ Tech Stack

Python

FastMCP (Model Context Protocol)

SQLite3

JSON / CSV

Datetime

👤 Author

Soumay Verma