On August 13, 2026, X published the source code for the algorithm that powers its default For You feed on GitHub under the Apache v2 license, roughly 10 to 15 times the size of its previous open source releases. The release covers the core ranking engine that selects and orders posts, along with the model configuration, filters, and the weight parameters used to combine different ranking signals. Alongside the code, the company introduced an in-app transparency tool that lets users download a JSON file showing how X’s ranking systems have affected their account and posts over the past calendar month.
The combined move opens the company’s most consequential recommendation system to outside inspection for the first time, and gives individual users a way to check, at the post level, whether their content was filtered, demoted, or otherwise labeled by ranking logic they previously could not see.
What is being released on GitHub?
The repository contains the code that pulls posts, scores them, and assembles the For You feed. That includes the systems that filter out potentially problematic, rule-violating content, the ranker, and the score. Developers can run the ranker and scoring components outside X’s infrastructure. The company previewed the codebase to external researchers familiar with recommendation systems, who were able to train and run X’s Phoenix scoring system using the open source code, a milestone the company described as a first for its transparency work.
X is also accepting pull requests on the repository. Engineers at the company will review submitted updates and may incorporate them into the algorithm. Systems that use Grok to predict whether a post could violate a rule are intentionally excluded from the release. X said withholding those components is meant to prevent bad actors from reverse-engineering detection logic to flood the network with spam.
How can users check if they have been shadowbanned?
A new transparency tool is rolling out to an Under the Hood page in the app’s settings. Users who have posted 10 or more times in the past 30 days can download an aggregate stats file covering the previous calendar month. The file shows whether any labels were applied to the account or to individual posts, and how the ranking systems interacted with that content.
Non-technical users can drop the JSON file into an LLM of their choice, point it at X’s GitHub repository, and ask for an interpretation. X is launching the feature as a pilot to a test group of accounts that are at least one year old, with a broader rollout planned afterward.
Why is X making this change now?
The release is framed as a response to ongoing criticism about how X’s algorithm shapes political conversations, elections, and the spread of misinformation. The platform has hosted political figures and high-profile individuals for years and is currently owned by Elon Musk, who has said he supported President Trump’s election. Concerns about algorithmic bias predate the current ownership: before Musk’s acquisition, Republican members of Congress accused the then California-headquartered company of leaning too far left and of shadowbanning conservative posts, claims Twitter denied at the time.
X’s stated goal is to let anyone in the public assess how posts are distributed on the platform, vet whether the playing field is level, and critique the system so it can be improved. The company has pointed to other transparency features, including the crowdsourced fact-checking system Community Notes, as evidence of a broader push toward openness.
Has X become more or less transparent overall?
The picture is mixed. On algorithmic and product transparency, X is releasing more than it has before, with the For You engine now available for outside audit. On corporate transparency, however, the company has moved in the opposite direction since going private under Musk. Without SEC reporting requirements, X has published fewer metrics on users, growth, revenue, and government takedown requests than it did as a public company.
Monthly active user figures have since returned to public view following the company’s merger with SpaceX, but other operational data points remain less frequently reported than they were before the take-private transaction.
What is still not in the open source release?
The Grok-based systems that predict whether a post could be violating a rule are kept proprietary, for the stated reason of limiting adversarial workarounds. The per-post score is not part of what external researchers received during the preview either. Researchers trained and ran the Phoenix scoring system themselves, but did not get access to the individual score assigned to each post.
FAQ
What did X release on August 13, 2026?
X released the source code for its For You feed ranking algorithm on GitHub under the Apache v2 license. The repository includes the core ranking engine, model configuration, filters, and the weight parameters used to combine ranking signals, and is roughly 10 to 15 times larger than X’s prior open source codebase.
How can a user check if they have been shadowbanned on X?
Users who have posted at least 10 times in the past month can open the new Under the Hood page in the app’s settings and download a JSON file showing any labels applied to their account or posts over the previous calendar month. Non-technical users can feed that file into an LLM along with X’s GitHub repo to interpret the results. The feature is initially limited to a pilot group of accounts that are at least one year old.
Why is X open sourcing its ranking algorithm?
The company said the release is meant to let outside researchers, developers, and the general public audit how posts are distributed on the platform, verify that the system is fair, and submit code improvements via pull requests. The stated goal is to make the algorithm visible to, and increasingly shaped by, the public.
This article summarizes reporting from techcrunch.com. See our editorial disclaimer for how our articles are produced.
Run a free scan to see your AI Visibility Score, SEO rating, and local citation accuracy.