Is this possible?

Sorry for such a clickbait title, but I was having trouble condensing my thoughts into a quick bite.

I run a Twitter page where I post images found across the internet and I currently use Buffer to schedule these posts. I do know that Buffer has an API, which leads me to believe people can do some cool stuff with it.

In essence, what I do is I look for these images from websites, write a quick blurb for them (usually sourced from the website itself) and throw them into my Buffer backlog. What I'm wondering is whether or not this process can be automated to the point where an algorithm can take the most successful keywords from my existing tweets, use those keywords to scour the internet and essentially compile these posts for me. In an ideal world, I'd have a massive automated queue of posts and I'd just go through manually and approve / deny them.

Not that it's particularly difficult to do it manually, but I'm looking for a solution that can be scaled.

Is there anything similar to this already? Or does it even sound possible?