Facebook lets users customize Feed and Reels recommendations

A direct way to influence recommendations

Meta has announced the launch of Your Algorithm for Facebook, a feature intended to give users more direct influence over the recommendations displayed in their Feed and Facebook Reels. According to the company, people can describe their preferences in natural language instead of relying solely on the platform’s automated signals. They can indicate the subjects and types of content they want to encounter more often, as well as those they would rather see less frequently.

The tool is presented as an AI-powered customization feature. Meta says it is designed to make recommendations more relevant by letting users interact with the systems that rank and recommend posts. The company says changes made through the feature are reflected in recommendations in real time, meaning users should see adjustments during the same session after updating their preferences.

Feed, Reels and the post menu

Your Algorithm applies to both Facebook Feed and Facebook Reels. Users can access it from the three-dot menu attached to individual posts, which Meta has updated to make preference controls easier to find. From that menu, people can open Your Algorithm directly or select Interested and Not interested to register a reaction to a particular item.

According to Meta, the Interested and Not interested controls may also appear beneath posts. These options are meant to provide a quicker way to signal whether a recommendation matches a person’s preferences. The feature does not remove Facebook’s existing automated ranking system; rather, it adds a more explicit channel through which users can communicate with it.

Natural-language feedback

The central promise of Your Algorithm is that users can explain their preferences in their own words. Meta says its systems may ask follow-up questions to gather more detail about what a person liked or disliked. That additional feedback is intended to help the platform refine later recommendations.

The company says it is also using large language models more extensively in ranking systems. In Meta’s description, these models can take account of interests, context and intent expressed through natural language. This is the technical basis for allowing a person to tell the service what they want more or less of without having to depend only on individual likes, follows or dismissals.

• Users can specify topics they want to see more often.
• They can identify subjects or content they want to see less often.
• They can use Interested or Not interested on individual posts.
• Meta says recommendations can update in real time after feedback.

A broader Meta personalization effort

The Facebook rollout follows the earlier introduction of Your Algorithm on Instagram. TechCrunch reports that Instagram first launched the feature for Reels in December 2025, then extended its customization options to Feed and Explore in June. Meta has therefore been expanding the same approach across several services rather than introducing an entirely separate Facebook system.

Threads has also received a related tool called Dear Algo. That feature likewise allows people to influence the content recommended to them. Together, these launches reflect Meta’s wider effort to give users more visible control over experiences that are increasingly shaped by AI-based content discovery.

The move also places Facebook closer to services that let customers personalize recommendation systems around declared interests. For years, social media feeds have largely been shaped by inferred behavior: what people watch, click, share, follow or skip. Your Algorithm adds stated preferences to that model, while Meta continues to operate the ranking and recommendation infrastructure itself.

What remains in Meta’s hands

The available information does not establish that users can fully redesign Facebook’s algorithm or inspect every factor used to rank content. Meta describes the feature as a way to influence recommendations, not as a mechanism for taking complete control of Feed or Reels. The company’s claims about relevance, real-time changes and improved understanding are descriptions of how it says the tools work.

The feature also depends on the choices people make while using it. A user who provides detailed feedback may give Meta more information about current interests, while someone who uses only Interested or Not interested may provide a narrower signal. Meta says follow-up questions can help make that feedback more specific, but it has not detailed every question users might receive or every ranking change that could follow.

For organizations and businesses serving English-speaking audiences, the practical change is that distribution on Facebook may be influenced more visibly by audience-declared interests. Publishers, brands and community groups will still need to produce content people choose to engage with, but they may also need to consider that users can now actively reduce or increase exposure to particular subjects in Feed and Reels.