Happy Holidays! 🎄
YouTube hosts long-form podcasts, announcements, and other primarily audio-driven announcements. However, unless you pay for premium, you can’t close the app. This isn't very pleasant. I’d love to listen to some podcasts on the go, but I don’t have enough time to watch a three-hour video. I really only want to get the summary and ask questions.
Today, we’re building a Chat with Youtube app using Streamlit, pytubefix, and OpenAI.
It’s pretty simple, and the app has less than 100 lines of code! All we need to do is:
Drop in the YouTube URL you’d like to process
Click “Process”
Ask questions!
Remember: You can get a step-by-step tutorial on this project using our AI tutor, PrompterGPT. Simply choose which project you’d like to work on, follow the instructions, and tell it what you want changed!
Application Steps:
Drop in a URL of a YouTube video that you would like to process.
Click “Process Video”. We will extract the audio from the YouTube video and create a transcription using OpenAI’s Whisper API.
Ask Questions! Ask any question about the video and receive and answer directly in Streamlit.
App Demo:
Tech Stack:
pytubefix
OpenAI’s Whisper
OpenAI’s GPT-4o-mini
Streamlit
Technical Workflow:
Sounds good? Let’s get to work. Click the button below to start building. PrompterGPT will walk you through the project step-by-step and help you troubleshoot, customize, or edit its scope. Just ask away!
Happy building, and see you next time.
— Nishad from Prompter
P.S. This project doesn’t interest you? Check out our previous project on how to build your own Scrum Master Check-in Tool
P.P.S for more daily thoughts on AI, and sneak peak of projects before they’re sent, make sure to follow us on X!