The problem with YouTube Kids is it filters channels, not videos, so I built something that scores every video individually with AI
YouTube Kids has a fundamental design flaw that nobody seems to talk about.
It filters at the channel level. Parents approve a channel, and everything that channel ever posts gets through. But channels aren't consistent. A channel that posts great educational content 90% of the time will occasionally post something you wouldn't choose for your kid. That one video still reaches them.
The obvious fix is to evaluate content at the video level, not the channel level. Every video individually scored before a child sees it.
That's what I built WayTube. AI scores each video across safety, language quality, educational value, and age appropriateness. Parents set a threshold once, the app filters automatically.
A few things that were more interesting than expected during the build:
The trust gap is real. Parents are instinctively skeptical of AI making content decisions for their kids. The question isn't whether the AI is accurate, it's whether parents believe it is. Building visible scoring (showing the score on each video) helped more than any explanation.
Google Play is harder than App Store for children's apps. Google's Families Policy has a content aggregation clause that's genuinely ambiguous for apps that curate third-party content. Three rejections so far, currently appealing with competitor evidence showing similar apps are approved.
The scoring dimensions matter more than the overall score. A video can score 75 overall but score differently across safety vs educational value vs age fit. Parents actually want to understand why a video was filtered, not just that it was.
Currently live on iOS. Android pending.
Happy to discuss the idea, the market, or the technical approach — especially the AI scoring architecture if anyone is curious.