Technology

YouTube Will Let Viewers Build Custom Feeds From a Prompt

Martin HollowayPublished 2w ago3 min readBased on 4 sources
Reading level
YouTube Will Let Viewers Build Custom Feeds From a Prompt
source:google.com

YouTube will let people in the US create their own video feeds by typing a plain-language request, on the web, on phones, and on TVs.

The option will sit on the homepage. Users choose "Your custom feed" from the options at the top of the Home tab and type a description of the feed they want, according to YouTube support documentation. The result appears as a separate feed, opened from a tab on the homepage. Users in the US will be able to try the system soon on all three screens, Engadget reported on Sept. 23.

The idea has grown since early tests. YouTube has expanded the option so people can make multiple personal feeds instead of one filtered view, SocialMediaToday reported in August. The prompt, the written request that defines the feed, can be edited at any time in the text box at the top of the custom feed to create a new feed, 9to5Google reported in May.

Comments get GIFs and synthetic voice

YouTube will soon let viewers post a GIF as a comment under a video, picked from a searchable GIF library. GIF comments will arrive in the mobile app worldwide sometime this fall.

The TV app is handling comments differently. It now lets viewers post voice comments using a text-to-speech tool, which turns typed text into spoken audio, on TVs that support it. Typing with a TV remote is slow. Speaking a comment from the couch lowers the effort, without needing a paired phone.

The broader question here is moderation. Text, GIFs, and synthetic voice can each be misused in different ways, so they need different safeguards. A searchable library suggests YouTube will curate the choices and set limits, rather than allow open uploads, which keeps the problem more contained.

Assistants, shows and live translation

YouTube is adding shopping tools to Ask YouTube, its video assistant, including video recommendations and a comparison table when people ask for product reviews. A request for reviews returns videos plus a simple side-by-side comparison.

Ask Music on YouTube Music will let listeners learn more about an artist and change their song queue in the same conversation.

YouTube is also introducing Podcast Lineup, a feature that uses generative AI, AI that creates new text and audio from patterns it has learned, to make a short weekly preview of suggested podcasts. It is a short sample, not a full transcript or an automatic playlist.

For creators, YouTube has extended its shows feature to Shorts. Creators can group Shorts into episodes with custom thumbnails and play them in order. It applies the episode format long used in television to vertical short video.

YouTube will start testing a Live auto-dubbing tool in early 2027, first for English and Spanish. Live dubbing is harder than dubbing a finished video, because delay, telling speakers apart, and translation accuracy all have to work in real time. Starting with two widely used languages keeps the early test simpler.

Separately, YouTube extended its exclusive Coachella streaming rights through 2030.

The broader context here is familiar from earlier recommendation systems. The main Home feed is tuned to keep the overall audience watching. Custom feeds hand more control to the viewer and split that one tuning task into many smaller ones defined by prompts. That can help when viewing has a clear purpose, such as learning a skill or following a niche hobby, while making results harder to measure. There is no longer one ranking to adjust.

In my view, the trade to watch is personalization against shared context. I raised two children through the PC era into mobile and streaming, and the pattern was consistent. The more control they had over their feeds, the faster they found what they liked and the narrower what they saw became. Prompt-made feeds will likely do the same for YouTube viewers. The tools are genuinely useful. They also leave it more to the viewer to look for variety.

Looking at what this means for builders, the near-term work is unglamorous. Understanding prompts for feeds, managing the GIF library, making text-to-speech reliable across TV models, and keeping live translation fast all have to work at YouTube scale before the features matter. If they do, what this enables is clear. Viewers get more precise control over discovery, creators get new ways to package Shorts and reach live audiences, and conversational assistants handle more of the work in between.