Inspiration
Taking notes is hard. Some lectures take up to 3 hours, and when there are no notes or slides provided, it is difficult to go through the lecture and study the material.
What it does
By simply uploading or providing a link to a video, our application can
- Generate a summary of the video with sections
- Make notes of the most important parts of the video
- Create a knowledge graph for better visualization of the information presented
How we built it
Given a video URL, we will download the video, extract the audio from it, and utilize AI to convert it into a textual transcript. Then the textual transcript is run through an LLM to generate summaries, notes, as well as knowledge graphs of the most important parts of the video. The final essence of the video is extracted and provided to the end user. The project is built on Streamlit and utilizes OpenAI to extract information from the video
Challenges we ran into
Downloading the video, extracting the audio, and then running through whisper took a large amount of time.
De-duplication of text transcript from audio
Accomplishments that we're proud of
Integration of Streamlit went smoothly with responsivity and interactivity.
Generating knowledge graph through the use of Open AI
What we learned
Many problems have existing open-source solutions. There is no need to reinvent the wheel
How to utilize OpenAIAPI
What's next for Learn.ai
Allow users to upload their own videos Enable multiple platforms
Built With
- openai
- python
- streamlit
Log in or sign up for Devpost to join the conversation.