Facebook users are getting a way to describe the content they want in their own words. Meta announced Your algorithm on October 13, adding written interests and more detailed feedback to Feed and Reels recommendations.

Recommendation systems normally infer interests from behavior, such as watched videos and reactions. The new control adds a direct statement from the person using the service. That can express a preference a simple like or dislike does not fully explain.

Meta says the tool is available through the three-dot menu, where people can add and manage topics. It also describes follow-up questions after someone selects Interested or Not interested, allowing more detail about the reaction.

Underneath those controls, large language models interpret the context and intent of freeform text, according to the company. That information joins the other signals used to rank content.

The company says recommendations can respond within the same session. Its announcement does not provide an independent measurement of how consistently the system follows a written preference or how long the resulting change lasts.

The distinction is between giving feedback and directly choosing the complete feed. Meta describes an additional input to its ranking system. It does not say that a written preference replaces all the system's other signals.

The change makes the recommendation process more explicit: a user can state an interest instead of waiting for the system to infer it. The algorithm still ranks the results.