ClipFinder

Logo
Overall view
Usage of ask function
Usage of summary function

Inspiration

Our team was tired of wasting time doing research watching several YouTube videos to only find the right information in the last one of them. It is hard to know if a YouTube video will be pertinent to our research.

What it does

We decided to create a Chrome extension that helps users ask questions about the video, find their answers quickly, and get a summary to cut down wasted viewing time. When on a YouTube video upon clicking on the extension the user can ask a question in the text box, or click on summarize to summarize the video. If an answer is found to the user's prompt, the extension responds with the answer and a timestamp at which the answer could be found.

How we built it

We have an HTML, CSS, and JavaScript frontend, using a Python Flask Server to link the JavaScript and Python sides of things. We used the YouTube API to get the transcript of the video which we fed to the OpenAI API, whose response we matched back with the transcript to find a timestamp. This response is then sent back to the front end via the server.

Challenges we ran into

Setting up the OpenAI tokens was a bit difficult because at times the documentation seemed insufficient. It was also our first time using the Flask server to connect a Python backend with a JavaScript frontend.

Accomplishments that we're proud of

Creating an esthetic user-friendly extension that incorporates AI and that could be used by all users, especially students.

What we learned

We learned how to create a Chrome extension, and how to manipulate YouTube's and OpenAI's API. How to manage both JavaScript and Python with a Flask Server.

What's next for ClipFinder

Expanding on features such as bookmarks for users, improving the UI and AI interactions. Finding appropriate clips on any YouTube video, not just one that is prechosen. Optimizing everything to increase speed.