X made the source code of its "For You" timeline ranking algorithm publicly available on GitHub under an Apache v2 license on August 13, company officials said [1, 2]. The open-sourced code, named "Phoenix," includes detailed model configuration, filtering rules, the core ranking system, and weighting parameters used to score posts for each user [1, 2, 3].

Keith Coleman, X’s vice president of product, said users can access "the core ranking code that pulls posts and ranks them for any given user and assembles the feed," including the filtering systems that remove problematic content. "Some of those systems, like the ranker and the score, you can even run yourself outside the company," he added, calling the release an unprecedented level of transparency into the platform’s algorithm [1, 3].

The Phoenix algorithm predicts engagement types — such as shares, replies, and follows — and assigns different weights to score posts. Negative actions like reports, mutes, "not interested" flags, and blocks carry heavy negative weights to reduce visibility, with reports weighted 468 times negatively compared to predicted likes [2]. Labels used to limit visibility include NSFW, spam, violence, impersonation, hateful conduct, and an incident-response label for emergencies [3].

Alongside the code release, X launched "Under the Hood," a transparency tool that allows users with at least 10 posts in the past 30 days to download a JSON file with aggregate, account-level data on labels affecting their account and posts [1, 3]. The tool is in pilot testing with a small group of accounts active for at least one year before a wider rollout [1, 3]. However, it does not provide data on individual posts, only monthly aggregates on visibility-impacting labels [3].

In an official statement, X said the goal is clear: "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" [2].

The pilot rollout of the "Under the Hood" transparency tool continues as X monitors feedback before expanding access to more users.