README.md

---
title: OmniTech Customer Support
emoji: 🤖
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: "5.9.1"
app_file: app.py
pinned: false
license: mit
python_version: "3.12"
---

# OmniTech Customer Support Chatbot

An AI-powered customer support chatbot built with:
- **Gradio** for the web interface
- **MCP (Model Context Protocol)** for tool orchestration
- **ChromaDB** for vector-based document search (RAG)
- **HuggingFace Inference API** for LLM responses

## ⚠️ IMPORTANT: Required Setup

**This Space requires `HF_TOKEN` to be set as a secret, or the LLM won't work!**

Without `HF_TOKEN`, you'll see truncated responses like:
> "Based on our Return policies: OmniTech Solutions Consumer Returns..."

**How to fix:**
1. On the Hugging Face spaces site (huggingface.co/spaces/userid/aia-capstone) click on the **Settings** link in the upper right. 
2. Scroll down until you see the **Variables and secrets** section.
3. Click **New secret** button on right.
4. Fill in the fields with Name: `HF_TOKEN`and Value: Your HuggingFace API token.
5. Click **Save** and the Space will restart automatically
6. You can watch progress back on the **App** tab.

## Features

- 🔍 **RAG-powered responses** - Answers based on product documentation
- 🎫 **Ticket creation** - Automatically creates support tickets when needed
- 👤 **Customer context** - Looks up customer information for personalized responses
- 📊 **Developer mode** - Toggle to see debug info, MCP stats, and knowledge search

## Usage

1. **Set HF_TOKEN first** (see above)
2. Select a customer from the dropdown
3. Type your question in the chat
4. Toggle "Developer Mode" to see debug information

## Support Categories

- Account Security (passwords, 2FA, account recovery)
- Device Troubleshooting (technical issues, setup)
- Shipping & Delivery (order tracking, shipping info)
- Returns & Refunds (return policy, warranties)

---
*Built for the Enterprise AI Accelerator Capstone*