All NewsEducationTV
Equities & FundsCrypto & Digital AssetsAI & TechnologyBusiness & CorporateUS Politics & PolicyGeopolitics & Global RiskMacro, Rates & FXCommodities & EnergyEuropean Politics & MarketsAsia-PacificReal Estate & Property
Story archiveAll categories
← All Stories

X opens source its ranking algorithm, enabling users to check for 'shadowbans'

Created at 13 Aug · 4:31 PM1 source↑ Market-relevant
IN SHORT

X is releasing the source code for its 'For You' timeline algorithm and core ranking engine on GitHub. The social network is also introducing a transparency tool allowing users to see if their accounts or posts have been affected by ranking systems, addressing concerns about content distribution and potential 'shadowbanning'.

✉Newsletter

PiQ Daily

Pick your topics. Get only what matters, on your cadence.

Key Numbers

10 to 15 timeslarger codebase size increase
10 or moreposts required for user stats download
1 yearminimum account age for pilot program

Who's Involved

X
social network expanding open source efforts
Keith Coleman
X's VP of Product
President Trump
elected with assistance from X's owner
X opens source its ranking algorithm, enabling users to check for 'shadowbans'

↳ Why This Matters

This move by X represents a significant step towards algorithmic transparency in social media, potentially setting a new standard for how platforms disclose their content ranking mechanisms and empowering users to understand and critique how their content is presented.

Key facts

  • X is open-sourcing its 'For You' timeline algorithm and core ranking engine.
  • Users will be able to see if their accounts or posts have been impacted by ranking systems.
  • The source code is available on GitHub under the Apache v2 license.
  • A new transparency tool will allow users to download their aggregate stats as a JSON file.
  • The tool will initially be a pilot for a test group of older accounts.
  • Some systems, like those using Grok for rule violation prediction, are excluded from the release.
  • X, formerly Twitter, is significantly expanding its open source efforts by releasing the source code for its 'For You' timeline algorithm and core ranking engine on GitHub. This move aims to increase transparency and allow users to understand how content is ranked and distributed on the platform.

    In addition to the code repository, X is introducing a new transparency tool within the app's settings. This feature will enable users who have posted at least ten times in the past month to download their aggregate statistics as a JSON file. This file will indicate if any labels have been applied to their account or posts, addressing long-standing concerns about 'shadowbanning' and algorithmic bias.

    Keith Coleman, X's VP of Product, stated that the company is releasing the core ranking code that assembles user feeds and systems that filter rule-violating content. He noted that some of these systems, like the ranker and score calculation, can even be run independently by users outside the company. Coleman expressed enthusiasm for the possibility of external developers contributing code improvements via GitHub pull requests.

    However, certain systems, such as those employing Grok for predicting rule violations, are not included in this release. X cited the need to protect against bad actors who might exploit such information to circumvent platform rules and flood the network with spam. The company hopes this increased transparency will allow the public to audit and critique the platform's distribution systems, fostering trust and a sense of fairness.

    This initiative comes amid ongoing scrutiny of X's algorithm's influence on political discourse and the spread of misinformation. While X is enhancing transparency in areas like its algorithm and Community Notes fact-checking, it has also become less forthcoming in other metrics since becoming a private company, no longer required to report to the SEC.

    Frequently asked questions

    X is open sourcing the source code for its 'For You' timeline algorithm and its core ranking engine. This includes parameters used to weight different signals for post display.

    This tool allows users who have posted at least ten times in the past month to download their aggregate statistics as a JSON file. This file will show if any labels have been applied to their account or posts.

    Systems that use Grok to predict rule violations are excluded to prevent bad actors from using the information to circumvent rules and flood the network with spam.

    Yes, developers can submit updates via pull requests to the GitHub repository, which X engineers will consider incorporating into the algorithm.

    What Happens Next

    01The transparency tool will roll out more broadly after a pilot phase.
    02X engineers will review and consider user-submitted code updates for the algorithm.
    03Users will be able to run parts of the ranking system outside of X.

    Get the newsletter.

    Pick the topics you actually care about. We'll email when there's news worth your time, on the cadence you choose. Cancel any time from your account.

    Cadence

    How It Developed

    X is expanding its open source codebase to include its 'For You' algorithm and core ranking engine.
    The company is making the source code available on GitHub under the Apache v2 license.
    X is adding a transparency tool to its app settings, allowing users to check if their content has been impacted by ranking systems.
    This new tool will enable users to download aggregate statistics as a JSON file.
    The feature will initially be available to a test group of accounts at least a year old.
    External researchers have already been able to run parts of the ranking code outside of X.
    X engineers will consider user-submitted code updates via pull requests to the GitHub repository.
    Systems using Grok to predict rule violations are not included in the release to prevent misuse.

    Sources

    T1
    X open sources its ranking algorithm, letting users see if they’ve been ‘shadowbanned’TechCrunch

    Related Stories

    DeepSeek releases V4 Pro AI model with enhanced capabilities and revised pricing
    13 Aug · 3:11 PM
    AI watermark debate intensifies after 'Made with AI' label appears on Karoline Leavitt's X post
    13 Aug · 2:11 AM
    DeepSeek to raise API prices for V4 models by up to 1100%
    13 Aug · 11:59 AM
    Claude users protest AI watermarking amid EU AI Act compliance
    12 Aug · 10:41 PM
    Flock Safety Enhances Privacy Measures Amid Misuse Backlash
    13 Aug · 5:06 PM