OhJeeves | Devpost

Professional business company
searching + overview feature
questions + summarization feature
questions + summarization feature (video)

Inspiration

fake news is bad!!

i wanted to create a way to help users avoid consuming fake news, without having to undergo the tedious effort of carefully reading every article they come across.

my solution seeks to not only tackle this problem, but in a way that is more realistic, also improving the current method that research is conducted.

Features

ask questions as you are reading an article, it is likely that you will encounter doubts. from localization errors, language barriers to something as simple as forgetting the content of the article as you read it, OhJeeves (jeeves) gives its users a way to clarify these doubts.

jeeves leverages natural language processing to gather context information both from the article and outside of it.

summarize articles jeeves reads articles before you, summarizing all key information into neat bullet-points, organized further under headers.

this makes it easier for when you just want to skim through an article, or just get the rough gist of it.

categorization even before you read an article, its important to know what you are getting into. you want to know whether the article even relates to what you are looking for, so that you don't waste your valuable time poking your nose into irrelevant sources.

jeeves categorizes all your searched articles into different categories. it breaks down the content of each article, and gives a percentage breakdown of how likely left-centre-right leaning a certain article is. Furthermore, with each search jeeves curates the results to give you an overall view of how the media is portraying different topics.

How we built it

currently, it is built using UiPath and its many features. it leverages various apis to work, ranging from voice to text for video summarizations, gpt-4 for summarizations, to google's gemini for better searching.