Ask and Answer Project

Overview

This is a customer service chatbot with Reason-and-Act multi-agent design addressing a wide range of customer pre- and post- purchase queries about products available for sale the on Canadian Tire Website.

Setup

1. Clone the repository

   git clone <your-repo-url>
   cd ask_and_answer

2. Install dependencies

   • Ensure you have Python 3.10+ and uv installed.
   • Install dependencies:

     uv sync

3. Environment Variables

   • Ensure all environment variables and certificate paths are set in .env file.

Running Connection Tests

To run the test scripts:

uv run test_embedding.py
uv run test_llm_agent.py

Requirements

• LLM: gpt-4o, gpt-4.1
• Embeddings: text-embedding-ada-002
• Framework: OpenAI Agents SDK
• Observability/Tracing: Langfuse
• Long-Term Memory: MongoDB
• Evals: TBD
• Demo: Gradio