YouTube launches Gemini-powered custom feeds that let you build your own recommendations

By AI123

YouTube announced on Wednesday, September 23, a new feature called custom feeds that lets users describe in their own words the kinds of videos they want recommended. Launched at the Made On YouTube event, the prompt-based tool uses Google's Gemini AI model to build personalized feeds, which are then pinned as a separate tab at the top of the YouTube home page.

Users can write detailed instructions for a feed, specifying topics to include or exclude, content priorities, and more. For example, YouTube suggests creating a feed of video podcasts for a 30-minute commute or requesting videos for a particular mood. The company said custom feeds will not replace the main recommendation feed; instead, users can jump between feeds when opening the app. The rollout across web and mobile begins in October 2026.

YouTube says custom feeds should help users better explore the platform's catalog of more than 20 billion videos. Emily Moxley, VP of Product Management for Viewer AI, said during a pre-brief with reporters that users are always one search away from a treasure trove of topics.

The move follows similar feed-building features from other social platforms, including Bluesky, Threads, Instagram, X, and Spotify, many of which are also AI-powered.

Explore topicsGeminiGoogleModelsYouTube论坛

Sources

View all