AI Agent Automation Hub — MCP-Based Multi-Tool Backend Platform

A production-quality personal expense tracker whose REST endpoints are also exposed as MCP (Model Context Protocol) tools, enabling AI agents (Claude, GPT-4) to add, query, and summarise expenses through natural language. Includes a LangChain agent layer and an n8n scheduled workflow for daily email summaries.

Built: April 2 – April 4, 2026
Author: Bharath Kumar Kalavakunta

---

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                      AI Agent Automation Hub                    │
│                                                                 │
│  ┌──────────────┐    ┌───────────────┐    ┌─────────────────┐  │
│  │  FastAPI App  │    │  MCP Server   │    │ LangChain Agent │  │
│  │  (Port 8000) │◄───│  (Port 8001)  │◄───│  (OpenAI/       │  │
│  │              │    │               │    │   Anthropic)    │  │
│  │  /api/...    │    │  add_expense  │    │                 │  │
│  │  CRUD + auth │    │  get_expenses │    │ "Add ₹500       │  │
│  └──────┬───────┘    │  get_summary  │    │  grocery"       │  │
│         │            │  delete_exp.  │    └─────────────────┘  │
│         │            └───────────────┘                         │
│  ┌──────▼───────┐                                              │
│  │  PostgreSQL  │    ┌───────────────┐    ┌─────────────────┐  │
│  │  (or SQLite) │    │  Django Admin  │    │  n8n Workflow   │  │
│  └──────────────┘    │  (Port 8080)  │    │  Daily Summary  │  │
│                      │  /admin/      │    │  → Email/Slack  │  │
│                      └───────────────┘    └─────────────────┘  │
└─────────────────────────────────────────────────────────────────┘

---

Project Structure

AI-Agent-Automation-Hub/
├── fastapi_app/            # Core REST API (FastAPI + SQLAlchemy)
│   ├── main.py             # App entry point, router registration
│   ├── models.py           # SQLAlchemy ORM models (User, Expense)
│   ├── schemas.py          # Pydantic request/response schemas
│   ├── database.py         # DB engine, session, init_db()
│   ├── auth.py             # JWT auth, password hashing
│   └── routers/
│       ├── expenses.py     # CRUD + summary endpoints
│       ├── users.py        # Register, login, /me
│       └── agent.py        # POST /api/agent/query
│
├── django_app/             # Django admin panel + auth
│   ├── manage.py
│   ├── config/             # Django project settings, URLs, WSGI
│   └── expenses/           # Django app: models, admin, views, URLs
│
├── mcp_server/             # MCP server wrapping FastAPI as tools
│   ├── server.py           # MCP tool definitions + server loop
│   └── example_client.py   # End-to-end demo client
│
├── agent/                  # LangChain agent layer
│   ├── agent.py            # AgentExecutor, OpenAI/Anthropic switch
│   └── tools.py            # LangChain Tool wrappers around REST API
│
├── n8n_workflow/           # n8n automation workflow
│   ├── workflow.json       # Importable n8n workflow definition
│   └── README.md           # Setup and usage guide
│
├── seed.py                 # Seed script — 18 sample expenses
├── requirements.txt        # All Python dependencies
├── .env.example            # Environment variable template
└── README.md               # This file

---

Quick Start

1. Clone and install dependencies

git clone https://github.com/kalavakuntabharathkumar/AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform.git
cd AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform

python -m venv venv
source venv/bin/activate        # Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Configure environment

cp .env.example .env
# Edit .env with your database URL and API keys

3. Start the FastAPI server

uvicorn fastapi_app.main:app --reload --port 8000

The API is now live at http://localhost:8000.
Interactive docs: http://localhost:8000/api/docs

4. Start the Django admin panel

# Apply migrations
python -m django_app.manage migrate --settings=django_app.config.settings

# Create a superuser
python -m django_app.manage createsuperuser --settings=django_app.config.settings

# Run Django dev server
python -m django_app.manage runserver 8080 --settings=django_app.config.settings

Django admin: http://localhost:8080/admin/
Expense dashboard: http://localhost:8080/dashboard/

5. Seed sample data

# Ensure the FastAPI server is running first
python seed.py

Adds 18 sample expenses across 8 categories. Demo credentials:

• Username: demo_user
• Password: demopassword123

---

API Reference

Authentication

Method	Endpoint	Description
POST	/api/auth/register	Create a new user account
POST	/api/auth/login	Login and receive a JWT token
GET	/api/auth/me	Get current user profile

All expense endpoints require Authorization: Bearer <token>.

Expenses

Method	Endpoint	Description
POST	/api/expenses/	Create a new expense
GET	/api/expenses/	List expenses (filterable)
GET	/api/expenses/summary	Spending summary by category
GET	/api/expenses/{id}	Get a single expense
PUT	/api/expenses/{id}	Update an expense
DELETE	/api/expenses/{id}	Delete an expense

Query parameters for GET /api/expenses/:

Parameter	Type	Description
category	string	Filter by category (partial match)
start_date	date	Start of date range (YYYY-MM-DD)
end_date	date	End of date range (YYYY-MM-DD)
limit	integer	Max results (default 50, max 200)

Query parameters for GET /api/expenses/summary:

Parameter	Type	Values
period	string	week | month | year | all

AI Agent

Method	Endpoint	Description
POST	/api/agent/query	Submit a natural language instruction

Request body:

{ "query": "How much did I spend on groceries this month?" }

---

MCP Tool Schemas

The MCP server (mcp_server/server.py) exposes four tools:

add_expense

{
  "name": "add_expense",
  "inputSchema": {
    "type": "object",
    "required": ["amount", "category"],
    "properties": {
      "amount":      { "type": "number",  "description": "Amount in INR (positive)" },
      "category":    { "type": "string",  "description": "Expense category" },
      "description": { "type": "string",  "description": "Optional note" },
      "date":        { "type": "string",  "format": "date", "description": "YYYY-MM-DD, defaults to today" }
    }
  }
}

get_expenses

{
  "name": "get_expenses",
  "inputSchema": {
    "type": "object",
    "properties": {
      "category":   { "type": "string" },
      "start_date": { "type": "string", "format": "date" },
      "end_date":   { "type": "string", "format": "date" },
      "limit":      { "type": "integer", "minimum": 1, "maximum": 200 }
    }
  }
}

get_summary

{
  "name": "get_summary",
  "inputSchema": {
    "type": "object",
    "properties": {
      "period": { "type": "string", "enum": ["week", "month", "year", "all"] }
    }
  }
}

delete_expense

{
  "name": "delete_expense",
  "inputSchema": {
    "type": "object",
    "required": ["expense_id"],
    "properties": {
      "expense_id": { "type": "integer" }
    }
  }
}

---

LangChain Agent — Example Commands

Natural Language Input	Expected Response
Add ₹500 grocery expense	"Added ₹500 expense under 'Groceries' ✅ ID: 21, Date: 2026-04-03"
How much did I spend this month?	"Spending summary for this month: Total ₹8,456.50 across 17 transactions..."
Show all transport expenses	Lists all expenses in the Transport category
Delete expense #5	"Expense #5 deleted successfully. ✅"
What's my biggest spending category this week?	Calls get_summary with period=week and interprets results

Running the agent via CLI

python -m agent.agent "Add ₹750 food expense — dinner at a restaurant"
python -m agent.agent "How much did I spend on transport this week?"
python -m agent.agent "Show my last 5 expenses"

Running via the REST endpoint

TOKEN="your_jwt_token_here"

curl -X POST http://localhost:8000/api/agent/query \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"query": "Summarise my expenses for this month"}'

---

Switching AI Providers

Set AI_PROVIDER in your .env file:

# Use OpenAI (default)
AI_PROVIDER=openai
OPENAI_API_KEY=sk-...
OPENAI_MODEL=gpt-4o-mini

# Use Anthropic Claude
AI_PROVIDER=anthropic
ANTHROPIC_API_KEY=sk-ant-...
ANTHROPIC_MODEL=claude-3-5-haiku-20241022

---

Deploying to AWS EC2 (Ubuntu)

1. Launch EC2 Instance

• AMI: Ubuntu 24.04 LTS
• Instance type: t3.small or larger
• Security group: open ports 22 (SSH), 8000 (FastAPI), 8080 (Django), 8001 (MCP)

2. Install Dependencies

sudo apt update && sudo apt upgrade -y
sudo apt install -y python3.11 python3.11-venv python3-pip postgresql postgresql-contrib

# Create DB
sudo -u postgres psql -c "CREATE USER expenseuser WITH PASSWORD 'yourpassword';"
sudo -u postgres psql -c "CREATE DATABASE expense_tracker OWNER expenseuser;"

# Clone and set up
git clone https://github.com/kalavakuntabharathkumar/AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform.git
cd AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform
python3.11 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env && nano .env   # fill in DATABASE_URL, API keys

3. Run FastAPI with Gunicorn + Uvicorn

gunicorn fastapi_app.main:app \
  -w 4 \
  -k uvicorn.workers.UvicornWorker \
  --bind 0.0.0.0:8000 \
  --daemon \
  --access-logfile /var/log/expense_api.log

4. MCP Server as a systemd Service

Create /etc/systemd/system/mcp-server.service:

[Unit]
Description=Expense Tracker MCP Server
After=network.target

[Service]
User=ubuntu
WorkingDirectory=/home/ubuntu/AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform
Environment=PATH=/home/ubuntu/AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform/venv/bin
ExecStart=/home/ubuntu/AI-Agent-Automation-Hub-MCP-Based-Multi-Tool-Backend-Platform/venv/bin/python -m mcp_server.server
Restart=always

[Install]
WantedBy=multi-user.target

sudo systemctl daemon-reload
sudo systemctl enable mcp-server
sudo systemctl start mcp-server

5. Django Admin

python -m django_app.manage migrate --settings=django_app.config.settings
python -m django_app.manage collectstatic --settings=django_app.config.settings
gunicorn django_app.config.wsgi:application --bind 0.0.0.0:8080 --daemon

6. Security Hardening

# Configure UFW firewall
sudo ufw allow 22/tcp
sudo ufw allow 8000/tcp
sudo ufw allow 8080/tcp
sudo ufw enable

# Use Nginx as a reverse proxy (recommended for production)
sudo apt install nginx
# Configure /etc/nginx/sites-available/expense-tracker to proxy to ports 8000 and 8080

---

Environment Variables Reference

Variable	Required	Description
DATABASE_URL	Yes	PostgreSQL or SQLite connection string
DJANGO_SECRET_KEY	Yes	Django secret key
JWT_SECRET_KEY	Yes	FastAPI JWT signing key
AI_PROVIDER	Yes	openai or anthropic
OPENAI_API_KEY	If using OpenAI	OpenAI API key
ANTHROPIC_API_KEY	If using Anthropic	Anthropic API key
FASTAPI_BASE_URL	Yes	Base URL of FastAPI server for MCP/agent
N8N_WEBHOOK_URL	Optional	n8n webhook URL for automation triggers
NOTIFICATION_EMAIL	Optional	Email for daily summary reports

---

License

MIT License — free to use, modify, and distribute.