Healthy Toronto Agent

English | 한국어

Find healthy and affordable food in Toronto - no cooking required.

Demo

Interactive chat interface with restaurant recommendations

Tech Stack

• LangChain: RAG (Retrieval-Augmented Generation) framework
• OpenAI API: Large Language Model for natural language understanding
• Chroma: Vector database for semantic search
• Streamlit: Web interface for interactive chatbot
• Python: Core programming language with pandas for data processing

Features

• 🔍 Semantic search through Toronto restaurant database
• 💬 Conversational AI with memory and context
• 🌐 Clean web interface with real-time responses
• 📊 Vector-based retrieval for accurate restaurant recommendations
• 🔄 Session management with conversation history

Project Structure

healthy-toronto-agent/
├── ingest.py           # Create vector DB from CSV
├── agent.py            # Streamlit web app with LangChain RAG
├── data/
│   └── healthy_toronto_eat.csv  # Restaurant dataset
├── db/                 # Chroma vector DB (not in git)
├── assets/             # App demo screenshots
├── README.md           # This file
├── development-log.md  # Development challenges and learnings
└── requirements.txt    # Required packages

Setup

1. Install requirements:

pip install -r requirements.txt

2. Add your OpenAI API key to a .env file:

OPENAI_API_KEY=your_key_here

Usage

1. Create vector database (once):

python ingest.py

2. Run the Streamlit web app:

streamlit run agent.py

3. Ask questions like:
   • "Find healthy vegetarian restaurants downtown"
   • "What are affordable options near University of Toronto?"
   • "Show me places with gluten-free options"

For detailed development insights and challenges faced, see development-log.md.

Inspiration

Inspired by Aurora Li and Cole Bowden at AI Meetup Toronto by AICamp (May 2025).

---

🗂️ Dataset constructed manually using publicly available restaurant information (e.g., address, category, dietary options). No copyrighted content used.