YouTube announced a new custom feeds feature on Wednesday that effectively allows viewers to design their own content recommendations using natural language prompts. The system provides users with an input prompt box where they can describe precisely what kind of videos they want to see, whether that means focusing on a specific niche topic or assembling media for a particular daily routine.
For instance, a commuter can instruct the platform to build a dedicated lineup of video podcasts suited for a 30-minute train ride. Alternatively, someone looking to unwind in the evening can request relaxing commentary videos, and the platform will assemble a tailored section matching that exact mood.
Gemini AI Drives Feed Personalization
The feature relies on Google's Gemini AI model to parse user requests and construct the corresponding video list. Once created, the custom feed is pinned right at the top of the YouTube home page inside its own dedicated tab. Because the curation relies on generative artificial intelligence, users can submit detailed and lengthy instructions. Viewers can specify which creators or topics to prioritize, what type of content to exclude entirely, and how they want the selection sorted.
Navigating Over 20 Billion Videos
Platform executives explained that custom feeds are intended to make navigating the service's massive video repository much simpler. During a briefing held ahead of the Made On YouTube showcase on Wednesday, Emily Moxley, VP of Product Management for Viewer AI, pointed out that having the world at one's fingertips is a signature aspect of the platform. Moxley stated that with over 20 billion videos in the YouTube corpus, users remain a single search away from a vast repository of subjects.
Coexisting Alongside the Default Home Feed
The company clarified that these personalized sections are not replacing the primary YouTube home feed, which will continue to deliver standard recommendations. Instead, users opening the application can switch over to whichever specialized tab fits their immediate intent. Support for creating multiple custom feeds will start rolling out next month across both mobile and web interfaces.
Part of a Broader Industry Shift
YouTube's rollout aligns with a wider trend across digital networks offering users greater control over their algorithmic streams. Decentralized platform Bluesky popularized custom feeds years ago to set itself apart from rival services, later adding an AI utility called Attie to streamline feed assembly. Meta's Threads, Instagram, X, and Spotify have similarly introduced features allowing people to shape how automated recommendation engines deliver content.
