-
-
Data flow and architecture
Inspiration
Having to stop a youtube video to confirm offhand facts is annoying and time consuming, and spending that time writing a program to do it for us is much more efficient.
What it does
Gets a transcript of the YouTube video using DeepGram. Analyse this transcript to find important phrases and timestamp them back with the original transcript. Search wikipedia for the most relevant article for each phrase, and pol specific facts by taking properties from the accompanying wikidata graph. These properties are filtered based on relevance and sent back to the chrome extension, along with the title and summary of the article to be displayed.
How we built it
A chrome extension communicating with a local server in python that performs the NLP and timestamping
Challenges we ran into
Accomplishments that we're proud of
What we learned
Lots of web apis
What's next for Youtube wiki referencer
Opening up to citations for other, (more reputable) sources
Log in or sign up for Devpost to join the conversation.