📧 Email Triage Automation API

A production‑ready Flask + MongoDB application that classifies incoming emails using AI logic, applies deterministic decision rules, and stores results for analytics.
This project demonstrates full‑stack engineering, API design, and business‑focused automation — built to showcase professional polish and recruiter‑ready skills.

---

🚀 Features

• AI Classifier (Mock + Extensible)
  Classifies emails into support, sales, spam, or other with urgency levels (high, medium, low).
• Decision Logic Engine
  Applies smart rules to determine actions: escalate, auto‑reply, ignore, priority‑reply, etc.
• MongoDB Storage
  Persists every triaged email with metadata (sender, subject, body, intent, urgency, decision, timestamp).
• Analytics Endpoints
  Query recent emails, filter by decision, and aggregate counts by intent.
• Recruiter‑Ready Code Quality
  Modular structure (ai_classifier.py, decision_logic.py, storage.py, app.py) with environment variables managed via .env.

---

🛠 Tech Stack

• Backend Framework: Flask (Python)
• Database: MongoDB (local or Atlas, viewed via Compass)
• Environment Management: python‑dotenv
• Deployment: Waitress (stable WSGI server for Windows)
• Optional AI Integration: OpenAI API (extendable)

---

📂 Project Structure

email_triage_api/
│
├── .env                # Environment variables (API keys, DB URI)
├── app.py              # Flask API entry point
├── ai_classifier.py    # AI classification logic
├── decision_logic.py   # Decision rules engine
├── storage.py          # MongoDB operations
└── requirements.txt    # Python dependencies

---

⚙️ Setup Instructions

1. Clone the Repository

git clone https://github.com/yourusername/email-triage.git
cd email-triage

2. Install Dependencies

pip install -r requirements.txt

3. Configure Environment Variables

Create a .env file with:

MONGO_URI=mongodb://localhost:27017/email_triage
OPENAI_API_KEY=your_openai_key  # Optional for AI extension

4. Run the Application

python app.py

or with Waitress:

waitress-serve --port=5000 app:app

Example Workflow

1. POST an email to /triage-email:

json { "sender": "client@example.com", "subject": "System outage", "body": "Critical issue, production system is down!" }

2. Classifier Output:

json { "intent": "support", "urgency": "high" }

3. Decision Logic:

json { "decision": "escalate" }

4. MongoDB Storage:

json { "sender": "client@example.com", "subject": "System outage", "intent": "support", "urgency": "high", "decision": "escalate", "timestamp": "2026-01-05T23:28:26.917036" }

---

👨‍💻 Author Anuj Mundu
Motivated MCA student, full‑stack developer, and aspiring business technologist. Skilled in Python, Flask, MongoDB, React.js, Node.js, and AI/ML integration. Focused on building scalable applications with measurable business impact.