A sample project to demonstrate the capabilities of OpenAI in creating the custom ChatBot to assist the customers on the webpage.
A user might find it difficult to Navigate through loops of website links.
Leveraging the OpenAI's chatGPT, website content can be tokenized, and be converted to embeddings using openAI. This model will be able to answer questions on chat Format.
In this Jupyter Notebook, first python script scrapes through website using Beautiful Soup and downloads html contents, in this instance ulm.edu and generates 9533 text content file. Then it is tokenized and converted to embeddings.
Following is the output of such methodology:
Note: Refere to web-qa.pdf for in depth result.