X Open-Sources Its Reach Code — and a Tool to Check If Your Posts Are Restricted

X released more detail on how its algorithmic code determines the reach of each user's posts. The stated aim is transparency into how the system amplifies or demotes in-stream updates, and a response to concerns about manipulation of content distribution.

What Was Released

Two things.

An expanded GitHub listing. X open-sourced the code affecting a post's visibility in the For You timeline, including detail on content labels that can impact post visibility.

A restriction-label checking tool in testing. Users can see labels applied to their account or posts that might limit visibility, and download information on labels or tags assigned over the past month.

X's framing: "Today, we are taking another major step in our ongoing efforts to increase transparency. We're open-sourcing the code that affects a post's visibility in the For You timeline, and releasing a new tool that shows people labels applied to their account or posts that might limit visibility. The goal is straightforward — we want people to be able to answer for themselves whether a platform is limiting their reach, whether the system is fair, and why they see particular content."

X VP of Product Keith Coleman said users can enter the app's source code into any LLM, along with their account data, to work out what affects their own posts' visibility.

It Is Not Actually Simple

Unpacking the information is not a straightforward process and requires meaningful manual input. For users in X's creator revenue share program, though, it could be a worthwhile process for understanding issues affecting content performance.

Shadowbans deserve a note. Shadowbans are generally not applied the way most users think they are. Account-level and post-level restrictions, however, are genuinely applied based on a range of parameters. This update is an attempt to let users verify that distinction themselves.

There Is a Precedent

X's track record on algorithmic transparency is not strong.

In 2023, months after acquiring the platform, Elon Musk authorized publishing a version of the feed algorithm on GitHub, vowing to maintain it with regular updates so anyone could see how posts are ranked and displayed.

The data shared was only a partial listing, with various elements withheld. The code was never updated despite the promises, leaving many users frustrated at the apparent lack of transparency.

Musk has nonetheless continued citing it as a key marker of transparency, insisting X is the most open and honest platform regarding content ranking. Last month he again vowed to fully open-source X's codebase to win user trust — a claim that sits oddly against his position that this was already done three years ago.

For how X's policy shifts intersect with creator compensation, see Creator Payouts Shift From Views to Originality.

X said it is testing the individual post label and tag tool with a randomized group of eligible accounts at least one year old.

What Marketers Should Take From It

There is now a path to verify reach loss instead of assuming it. When a brand account's impressions fell, teams could only speculate. If the label tool expands, at minimum you can establish whether a restriction label was actually applied. That separates hypothesis from fact in social performance reporting.

But it is still a test. The rollout targets a randomized group of accounts at least a year old, brand accounts may not be included, and interpretation requires manual work. This is not ready to enter an operating process.

A transparency announcement and the actual disclosure are different things. The 2023 release was partial and never updated. Whether this one is maintained remains to be seen. That history is exactly why platform policy announcements should not become planning assumptions on their own.

Most shadowban claims are misreadings. The sense that reach has dropped usually reflects engagement rate on the content rather than an algorithmic penalty. Wider access to the label tool could settle much of that argument.