This is a side project inspired by HN comments. YouTube has great content (Mark Rober, Smarter Every Day, etc.), but also a lot of junk. YouTube Kids wasn’t any better—my 8-year-old kept getting pulled into endless autoplay loops. I don’t mind screen time, but I want it to be limited and worthwhile.
Some parents here mentioned carefully curating every video their kids watch. I thought, “that sounds like way too much work.” So naturally, I did the even harder thing and built an app for my kids :)
SelecTube is a hand-picked set of YouTube creators I’m comfortable with them watching. No autoplay, no ads (if you have YT Premium). It’s worked surprisingly well—my kids watch less overall, and what they do watch is usually educational.
I also used it as an excuse to play with AI tools. The whole project was built in Cursor, and the landing page is mostly AI-generated. As someone who codes infrequently, Cursor felt like a superpower—some things it couldn’t figure out, but also big accelerations.
Now I’m debating whether to leave it as a personal tool or keep building. Ideas on the list:
- Real design polish
- Accounts for watch history + parent customization
- Broader content library
- Smarter discovery that doesn’t recreate the algorithm problem
Would love feedback. Thanks for taking a look!
0 comments
—
Show HN: SelecTube – Curated YouTube for Kids · HackerTrans
Some parents here mentioned carefully curating every video their kids watch. I thought, “that sounds like way too much work.” So naturally, I did the even harder thing and built an app for my kids :)
SelecTube is a hand-picked set of YouTube creators I’m comfortable with them watching. No autoplay, no ads (if you have YT Premium). It’s worked surprisingly well—my kids watch less overall, and what they do watch is usually educational.
I also used it as an excuse to play with AI tools. The whole project was built in Cursor, and the landing page is mostly AI-generated. As someone who codes infrequently, Cursor felt like a superpower—some things it couldn’t figure out, but also big accelerations.
Now I’m debating whether to leave it as a personal tool or keep building. Ideas on the list: - Real design polish - Accounts for watch history + parent customization - Broader content library - Smarter discovery that doesn’t recreate the algorithm problem
Would love feedback. Thanks for taking a look!