X open sources its ranking algorithm and adds a tool to check if your posts have been filtered

By NewSunSEO | August 14, 2026

X has released the source code that powers its For You feed on GitHub under an Apache 2.0 license and is rolling out a new transparency tool that lets users see whether their account or posts have been affected by any of the platform’s ranking systems. The release, announced on August 13, 2026, expands a previously narrow open-source effort into a codebase roughly 10 to 15 times larger than before, including model configuration, filters, and the core ranking engine.

What is now in the open source release?

The new GitHub repository contains the core ranking code that pulls posts, scores them for any given user, and assembles the feed. It also exposes the parameters used to weight different signals, which determines which posts are actually displayed. Internal engineers previously previewed the codebase to outside researchers familiar with recommendation systems, who were able to train and run the Phoenix scoring system using the open source code, a milestone that the company framed as important for its transparency effort.

How can users check if their posts were filtered?

A new transparency tool is being added to an Under the Hood page inside the app’s settings. Users who have posted at least 10 times in the past 30 days can download their aggregate ranking data as a JSON file. The file shows whether any labels have been applied to their account or posts during the past calendar month. Non-technical users can drop the file into an AI assistant of their choice, point it at X’s GitHub repository, and ask for an interpretation.

The tool is initially being piloted with a test group limited to accounts that are at least one year old, with a broader rollout planned later.

Can outside developers contribute to the algorithm?

Developers can submit pull requests to the GitHub repository, which X engineers will review for possible inclusion. Not every submission will be accepted, but the company has signaled that public contributions to the algorithm itself, not just visibility into it, are part of the plan. Keeping the ranker and score systems runnable outside the company is a deliberate choice meant to lower the barrier for independent auditing.

What is being left out of the release?

The systems that use Grok to predict whether a post could violate a rule are not part of the open source release. The company cited the need to keep bad actors from reverse-engineering its enforcement tools to spam the platform. Ranking signals tied to safety scoring therefore remain proprietary, even as the core feed assembly code is published.

Why is X pushing transparency now?

The release is framed as a response to long-standing concerns about how X’s algorithm shapes politics, elections, and the spread of misinformation. Critics in earlier years, including members of Congress, alleged that the platform leaned left and shadowbanned conservative voices, claims the prior owner consistently denied. The goal of the open code, according to the company, is to let any outside party audit the system, critique it, and flag cases where the playing field is not level.

What about openness elsewhere at X?

The new openness around the algorithm sits alongside other X transparency features, such as the Community Notes crowdsourced fact-checking system. Outside of product, X has become less open in other respects since going private. It is no longer required to file with the SEC, and it has reduced reporting on user metrics, growth, revenue, and government takedown requests. Now merged with SpaceX, X has resumed publishing monthly active user figures.

FAQ

What did X release on GitHub?

X released the source code for its For You feed under the Apache 2.0 license, including the core ranking engine, model configuration, filters, and the parameters used to weight ranking signals. The updated codebase is roughly 10 to 15 times larger than its previous open source footprint.

How can I check if my posts are being shadowbanned on X?

Open the Under the Hood page in the app’s settings and download your aggregate ranking data as a JSON file, which is available to accounts that have posted 10 or more times in the past 30 days. The file shows whether any labels have been applied to your account or posts in the past calendar month. You can then feed it into an AI assistant along with X’s GitHub repo to interpret what the data means.

Which parts of X’s algorithm are not being open sourced?

The Grok-based systems that predict whether a post could violate a rule are excluded from the release to keep bad actors from gaming content moderation. Everything else in the core For You ranking pipeline, including the Phoenix scoring system, is included, with outside researchers already able to train and run that scoring system using the published code.