TLDW - Too Long; Didn't Watch

Curious minds are fueled by learning and exploration, but it can be hard to find the time for that. TLDW allows you to dive into any and every topic by turning videos into organized and concise notes.

Comment

Inspiration

Today, anything can be learned on the internet with just a few clicks. Information is accessible anywhere and everywhere- one great resource being Youtube videos. However accessibility doesn't mean that our busy lives don't get in the way of our quest for learning.

TLDR: Some videos are too long, and so we didn't watch them.

What it does

TLDW - Too Long; Didn't Watch is a simple and convenient web application that turns Youtube and user-uploaded videos into condensed notes categorized by definition, core concept, example and points. It saves you time by turning long-form educational content into organized and digestible text so you can learn smarter, not harder.

How we built it

First, our program either takes in a youtube link and converts it into an MP3 file or prompts the user to upload their own MP3 file. Next, the audio file is transcribed with Assembly AI's transcription API. The text transcription is then fed into Co:here's Generate, then Classify, then Generate again to summarize the text, organize by type of point (main concept, point, example, definition), and extract key terms. The processed notes are then displayed on the website and coded onto a PDF file downloadable by user. The Python backend built with Django is connected to a ReactJS frontend for an optimal user experience.

Challenges we ran into

Manipulating Co:here's NLP APIs to generate good responses was certainly our biggest challenge. With a lot of experimentation (and exploration) and finding patterns in our countless test runs, we were able to develop an effective note generator. We also had trouble integrating the many parts as it was our first time working with so many different APIs, languages, and frameworks.

Accomplishments that we're proud of

Our greatest accomplishment and challenge. The TLDW team is proud of the smooth integration of the different APIs, languages and frameworks that ultimately permitted us to run our MP3 file through many different processes and coding languages Javascript and Python to our final PDF product.

What we learned

Being the 1st or 2nd Hackathon of our First-year University student team, the TLDW team learned a fortune of technical knowledge, and what it means to work in a team. While every member tackled an unfamiliar API, language or framework, we also learned the importance of communication. Helping your team members understand your own work is how the bigger picture of TLDW comes to fruition.

What's next for TLDW - Too Long; Didn't Watch

Currently TLDW generates a useful PDF of condensed notes in the same order as the video. For future growth, TLDW hopes to grow to become a platform that provides students with more tools to work smarter, not harder. Providing a flashcard option to test the user on generated definitions, and ultimately using the Co-here Api to also read out questions based on generated provided examples and points.

Built With

Share this project:

Updates