Hi, everyone. I just made a new Python-based automation system that makes blog articles. So, essentially, it scrapes the top search results from Google based on a keyword and then passes these keywords to Gemini AI, which in turn gives a title for an article. Then, it searches for that article on YouTube and gets the transcription of the most recent top video. And again, that is given to Gemini Ai which generates a table of contents (TOC). Then, the program processes that TOC and passes each heading to Gemini Ai one by one with (a lot of) guidelines and the transcription we just extracted. So, after everything's done, it creates an AI image prompt for the article and then an AI image using Flux Schnell, and finally everything is uploaded to Firebase's Firestore DB.
Then, that data is accessed in the front end of my new Next.js blog and all that good stuff is done programmatically.
After all that I also have a make.com automation which triggers and uploads these generated images and article links to Linkedin, Instagram, and Facebook. All this system completely runs on auto pilot without any input from my side.
For reference, here is a blog post generated using the system:-