Latest Liverpool FC News
LFCFeed.com
Site: LFCFeed.com
Remixed GunnerNews.com Replit
Had to get Assistant to rewrite replit.md to be for Liverpool, not Arsenal, because it was reusing the GunnerNews.com replit.md file and kept rewriting the code for Arsenal
It reused the news sources fine and switched the teams seamlessly
Had to provide it the LFC logo to replace the Arsenal logo (it made up its own SVG file at the beginning)
Added current table position (premier league and champions league) - scraped Premier League and Champions league sites. This took building the component first with fake data to make sure it looked right. Then a few rounds to simplify the data source, acquisition and parsing logic to get exactly what we needed. Then had to redo the data scraping with BBC site for both tournaments, which was fine
Added latest match/result and next match. Scraped LiveSoccerTV.com and provided TV/stream channel data based on user location. Ultimately had to scrap this because we were being blocked by LiveSoccerTV, and got the info directly from the official Liverpool FC site instead.
I’d initially envisioned having 3 separate components for the Premier League, Champions League and Match data. But after seeing the initial layout and working to reduce the amount of space taken, I realized the three components could be combined more succinctly as 3 subsections within a Standings & Matches component.
Asked Replit to make the red in the font and other elements to be more harmonious with the Liverpool red, which it did rather nicely without making it all the same red.
Also got Replit Agent to simplify the process to get actual favicon for news articles without guessing the domains. It had to ignore the Google RSS redirects and get to the source article link to extract the favicons
Launched
August 134 2025—one day before Liverpool kicks off the 2025–26 Premier League season and the defense of their championship.
Learnings
Using Assistant is expensive. I thought it would be better in figuring out scraping and parsing data for the standings and matches but not before burning through credits. Agent worked just fine
One problem at a time. One data source problem at a time. Then handle style and issues separately.
Have AI update replit.md with successes so we can avoid rehashing problems later.
Forgot and had to create new YouTube API key instead of reusing the one from GunnerNews.com. We were hitting API quota limits too soon. Then had to update in secrets. Also remembered to run a security check, vulnerabilities were found and addressed.