YouTube has introduced a feature called "custom feeds" that lets users write natural-language prompts describing the kinds of videos they want, and then builds a dedicated feed around that description.
To create a feed, users type a prompt into a box on YouTube specifying the feed’s purpose or theme. YouTube suggested examples such as a collection of video podcasts for a 30-minute train commute or a set of videos matching a particular mood, like "relaxing commentary videos to unwind with."
The feature uses Google’s Gemini AI model to interpret the prompt and assemble a feed according to the user’s instructions, including details about what to prioritize or exclude. Once created, the custom feed is pinned as its own tab at the top of the YouTube home page.
This move follows a broader industry trend where platforms give users tools to shape their own recommendation experience. Bluesky helped popularize the concept and has rolled out an AI tool called Attie; other services, including Meta’s Threads, Instagram, X, and Spotify, have introduced similar, often AI-driven feed-building features.
YouTube frames custom feeds as a way to navigate its extensive content library. Emily Moxley, Vice President of Product Management for Viewer AI at YouTube, noted that the platform hosts more than 20 billion videos, making targeted ways to explore content valuable.
The custom feeds are not intended to replace YouTube’s main recommendation feed; rather, they provide separate, purpose-built tabs so users can quickly access the feed that best fits their needs when they open the app.
YouTube said support for creating multiple custom feeds will roll out on web and mobile starting next month.
