YouTube Video Suggester

The Problem

The world's largest learning platform is optimised against you.

YouTube is the largest learning platform in the world, but its recommendation algorithm is optimised for engagement, not learning outcomes. When you search for a topic — say "prompt engineering for product managers" — you get a mix of beginner overviews, clickbait thumbnails, and content that is two years out of date.

Finding a video that matches your current knowledge level and specific goal can take 20–30 minutes of trial and error: click, watch 3 minutes, realise it's too basic, go back, repeat. That friction cost compounds across every topic you want to learn.

The platform gives you no way to express intent beyond a search query. You can't say "I already know the basics, show me advanced content from credible creators under 20 minutes." The algorithm has no model of you as a learner — only of you as a viewer to be retained.

The people who suffer most from this are professionals with a specific skill gap and limited time — exactly the people who most need high-quality, targeted learning content.

The Solution

Describe your goal. Get exactly the right video.

An AI-powered video suggestion tool where you describe your learning goal in plain language — including your current level and what you want to achieve — and the tool returns a curated shortlist of YouTube videos matched to your intent. Uses semantic search ranked by relevance, creator credibility, and content freshness.

The interface is intentionally minimal: one input box, one results list. Each result shows the title, channel, duration, and a one-line reason why it matched your specific query. No thumbnails, no view counts, no engagement bait — just signal.

Key design decision: the user's stated level and goal are treated as hard constraints, not soft preferences. A beginner query and an advanced query for the same topic return fundamentally different results — the LLM interprets intent before the search runs.

Product Screenshots

Interface