Personally, YouTube has contributed more to my education and professional life than my Master’s or Bachelor’s degree. Having said that, sometimes I prefer the information in a distilled format for easier comprehension or while recalling content from a particular video.

YouTube Distillery accepts the URL for any YouTube video and summarizes its contents in a easy to comprehend manner. It relies on OpenAI’s GPT models to summarize the transcripts from the videos on YouTube.

• AWS sign up
• Serverless Framework
  • Lambda → layers add complexity
  • https://github.com/erenyasarkurt/OpenAI-AWS-Lambda-Layer/blob/main/README.md
• Python
  • requirements.txt
  • https://learnpython.com/blog/python-requirements-file/
• Youtube Transcript Api
  • Fetches either auto-generated or manually uploaded transcript from YouTube
• OpenAi API
  • GPT 3.5-turbo used for summarizing transcript

Challenges

• Lots of boilerplate infrastructure to get started
• synchronous API → Asynchronous API (requests take long to process)
  • add SQS