AI-driven conversational assistant for e-commerce workflows.
Built on Rasa 3.x for intent classification and dialogue management, with a React frontend and Node.js middleware.

• Order tracking via validated IDs
• Product and inventory lookup
• Return / refund flow automation
• Payment and shipping inquiries
• Context-aware multi-turn conversations

git clone <repo-url>
cd rasa-ecommerce-assistant

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

# Frontend setup
cd src && npm install && cd ..

# Train and launch
rasa train
rasa run --enable-api --port 5005 &
rasa run actions --port 5055 &
cd src && npm start

Intents: track_order, return_policy, product_search, shipping_info, payment_methods

Entities: order_id, product_name, shipping_type, payment_method

rasa-ecommerce-assistant/
├── actions/          # Custom Python actions
├── data/             # NLU data, stories, rules
├── models/           # Trained Rasa models
├── src/              # React client
├── config.yml        # Rasa pipeline & policies
├── domain.yml        # Intents, entities, slots, responses
├── endpoints.yml     # Action server configuration
└── requirements.txt  # Python dependencies

• Response latency < 200 ms
• Intent accuracy ≥ 95%
• ~50 concurrent sessions (test environment)

MIT © Contributors