I built this basically to learn RAG. Ended up getting ahead of myself and building a full scraping pipeline that runs hourly across the news source and embeds them into a vector dab for retrieval :)
Wow, it seems to do exactly what i wanted to build for me personally, so i can save the time, that i loose when doomscrolling the news.
i'd put in the questions, that i am really interested in, and have the news on that question on a daily basis or more often.
An E-Mail service to that questions would be good, too, so i'd not even have to visit the browser. An rss-feed on my questions would be great.
So, yes, i could build it myself, but i didn't find the time for it for now. Great achievement! But be cautious and rate-limit the site, because someone could drive up your openai-bills, or whatever AI you are currently using.
I think this would be great for individuals and self-hosted.
i'd put in the questions, that i am really interested in, and have the news on that question on a daily basis or more often.
An E-Mail service to that questions would be good, too, so i'd not even have to visit the browser. An rss-feed on my questions would be great.
So, yes, i could build it myself, but i didn't find the time for it for now. Great achievement! But be cautious and rate-limit the site, because someone could drive up your openai-bills, or whatever AI you are currently using.
I think this would be great for individuals and self-hosted.