Elyza Myanmar Chatbot

A rule-based Myanmar language chatbot with emotion detection, served as a Flask REST API.

Key Highlights

• ◆Emotion classification across 6 categories (Joy, Sadness, Love, Anger, Fear, Surprise)
• ◆Rule-based Burmese NLP response engine
• ◆Flask REST API with CORS support
• ◆Minimal HTML UI for browser testing

Overview

Elyza is a Burmese-language chatbot inspired by the world's first chatbot, ELIZA - reimagined for Myanmar language. It detects emotional context in user messages and returns empathetic, context-aware replies. Served as a Flask REST API with a lightweight HTML UI for testing. Designed for educational purposes, research, and NLP experimentation in Burmese.

Features

• –Rule-based Myanmar language response engine
• –Emotion detection across 6 emotions with confidence scores
• –Flask REST API with CORS enabled
• –Minimal HTML UI - open directly in any browser
• –Quit keyword detection (ဘိုင်, bye, exit, etc.)

API Endpoints

Installation & Setup

git clone https://github.com/sai-zack-dev/myanmar-chatbot.git
cd myanmar-chatbot

# macOS / Linux
python3 -m venv venv && source venv/bin/activate

# Windows
python -m venv venv && venv\Scripts\activate

pip install -r requirements.txt
# flask>=3.0.0, flask-cors>=4.0.0

python app.py
# API live at http://localhost:5005

Usage

# Request
curl -X POST http://localhost:5005/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "မင်္ဂလာပါ"}'

# Response
{
  "user_message": "မင်္ဂလာပါ",
  "bot_reply": "ဟုတ်ကဲ့၊ ဆက်ပြောပါဦး။",
  "is_quit": false,
  "emotion": {
    "label": "Joy",
    "confidence": 66.7,
    "scores": { "Sadness": 8.3, "Joy": 66.7, "Love": 8.3, ... }
  }
}

Related Projects