YouTube Lets Users Build Their Own Recommendation Feeds With AI

YouTube will let users describe, in their own words, the videos they want recommended to them. The company announced the feature, called custom feeds, on Wednesday at its Made On YouTube event.

Users type a description into a prompt box — for example, video podcasts for a 30-minute train commute, or "relaxing commentary videos to unwind with." The feature leverages Google's Gemini AI model to build the feed around the request, and the result is pinned to the top of the YouTube home page in its own tab. Because generation is AI-powered, prompts can be lengthy and detailed, specifying what the feed should feature, what to exclude, and what content to prioritize.

The launch lands in a market where user-built feeds have moved from niche experiment to mainstream product strategy. Bluesky began popularizing the idea years ago as a way to differentiate itself from rivals like X, and this year rolled out an AI tool, Attie, to make feed-building easier. Meta's Threads, Instagram, X and, more recently, Spotify have all launched their own feed-building features, many of them also AI-powered.

For YouTube, the stakes are scale. The platform now hosts over 20 billion videos, and custom feeds are the company's latest attempt to help users navigate that catalog. "One of the amazing things about being on YouTube is that the world is always at your fingertips," Emily Moxley, VP of Product Management for Viewer AI, told reporters in a pre-brief ahead of the event. "There's over 20 billion videos in the YouTube corpus, so you're always one search away from a treasure trove of topics."

The custom feeds will not replace YouTube's main recommendations feed. Instead, users will be able to jump straight into whichever feed suits their purpose when they open the app.

Support for creating multiple custom feeds will roll out across YouTube on web and mobile starting next month, the company said. The move signals that AI-generated personalization — where users prompt a feed into existence rather than accept an opaque ranking system — is becoming a baseline expectation across the largest content platforms.