X Ranking Algorithm: See What’s Really Happening

X Open-Sources Its “For You” Ranking Algorithm and Adds Shadowban Transparency Tools

X has expanded its open-source recommendation code and introduced new tools that let users check whether their posts or accounts may have been downranked or labeled by its visibility systems. The update centers on the “For You” feed, the platform’s default timeline, and aims to make ranking and moderation signals more transparent to users.

What X Released

The company made the source code for the “For You” timeline available on GitHub under the Apache v2 license, expanding on earlier transparency efforts. The release includes more detail about the model configuration, filtering logic, and core ranking system behind how posts are surfaced.

• Open-source code for the “For You” feed ranking system
• Additional model configuration and filter details
• Core ranking system information for how posts are scored

How the New Transparency Tool Works

X also introduced an “Under the Hood” page in settings that lets eligible users download aggregate stats as a JSON file. The tool is designed for accounts that have posted at least 10 times in the past month, and it shows whether labels have been applied to the account or its posts over the previous calendar month.

X says users can then examine that JSON data alongside the public code to better understand how the platform’s systems may have affected reach or visibility.

Why This Matters for Users

The update is significant because it gives creators, brands, and everyday users a clearer way to investigate whether reduced reach may be tied to platform labels or ranking decisions. That makes the long-running “shadowban” debate easier to evaluate using actual platform signals rather than speculation.

• Users can check for visibility-impacting labels
• Creators can better diagnose sudden drops in reach
• Researchers can inspect how ranking and filtering are implemented

What the Open-Sourced Code Reveals

The repository shows that X’s feed combines in-network content from followed accounts with out-of-network content discovered through machine-learning retrieval. The system then ranks content using a Grok-based transformer model, with filtering and labeling steps affecting what appears in the feed and how prominently it is shown.

SEO Keywords Used

Primary keyword: X open-sources For You ranking algorithm

Secondary keywords: shadowban transparency tools, X ranking system, downranked posts, For You feed algorithm, Under the Hood tool

Conclusion

X’s latest move combines open-source disclosure with user-facing transparency, giving people more visibility into how posts are ranked and whether account labels may be limiting distribution. For anyone watching platform algorithm changes, this is one of the most direct attempts yet to explain how feed visibility is shaped inside X.