Abstract:
YouTube Transcript Summarizer the project aims to automatically generate concise
summaries of YouTube video transcripts. It utilizes machine learning models to summarize
the text and integrates with the YouTube Data API V3 to fetch the video data.The front end of
the YouTube Transcript Summarizer developed with Streamlit provides a user-friendly interface
where users can input the YouTube video URL. Once the user submits the video information,
the application utilizes the YouTube Data API V3 to fetch the necessary data, including the
video transcript. The transcript is then passed through deep learning models to generate a
concise summary. In this system, two models Seq2Seq and Transformer-based model are
used to generate a summary and calculate TF-IDF vectors for the summaries and compute
the cosine similarity. Finally, it compares the cosine similarities to determine the best summary
and returns it. The system can send users an email containing the summary along with a link
to download the audio version.
This tool is particularly beneficial for individuals with disabilities, as it allows them to
focus on the most relevant information and aids those with hearing or vision difficulties.
By providing transcripts, YouTubers can reach a wider audience, and users can save time
by consuming summarized content. With the exponential growth of YouTube videos, a
summarizer becomes an invaluable resource to quickly grasp the main points of a video. It also
helps in developing summarization skills, benefiting various professions. Overall, a YouTube
Transcript Summarizer enhances time efficiency, writing abilities, and accessibility to YouTube
content
