X is releasing a much larger portion of the code behind its “For You” feed and introducing a tool that will show users whether ranking systems or content labels have affected their posts. The company says the expanded release includes its core ranking system, filters, model configuration, and the weights assigned to different engagement signals.
The code is available through X’s x-algorithm repository on GitHub under the Apache 2.0 license. The repository includes Phoenix, X’s recommendation system for retrieving and ranking posts, along with an end-to-end inference pipeline that developers can run outside the company.
X vice president of product Keith Coleman said the newly published codebase is roughly 10 to 15 times larger than earlier releases. It provides more detail on how posts are selected, scored, filtered, and assembled into a user’s For You feed.
Users Will Be Able to Check Ranking Labels
X is also rolling out an “Under the Hood” section in its settings that lets eligible users review whether labels or ranking systems have affected their accounts or posts. Users who have posted at least 10 times during the previous month will be able to download aggregated information as a JSON file.
The file will cover labels applied during the previous calendar month. Users can inspect the data themselves or provide it to an AI model alongside X’s public codebase to ask how particular labels may have influenced distribution.
The feature will initially be tested with a limited group of accounts that are at least one year old before expanding more broadly. X says the aim is to give users more direct evidence when questioning whether their posts have been restricted or ranked differently.
Some Safety Systems Remain Private
Not every part of X’s recommendation and moderation infrastructure is included. Systems that use Grok to assess whether posts may violate platform rules remain private because X says publishing those details could help spammers and other users bypass its protections.
External researchers previewed the expanded release before publication and were able to train and run the Phoenix scoring system using the available code. Developers can also submit proposed changes through GitHub pull requests, which X engineers may consider for incorporation into the live system.
The release follows years of scrutiny over how X and its predecessor, Twitter, rank political content, misinformation, and other posts. Coleman said X wants researchers and users to be able to audit and criticize the recommendation system and assess whether posts are distributed fairly.
X has previously released parts of its recommendation algorithm, but the latest publication exposes substantially more of the system used for its default feed. Its official GitHub organization lists the x-algorithm project as the code powering X’s For You feed.
Featured image credits: Wikimedia Commons
For more stories like it, click the +Follow button at the top of this page to follow us.