2025 Transparency Report Overview
January 29, 2026
by Aaron Rodericks
When we set out to build Bluesky, we started from a simple premise: social media is broken at its foundation. For too long, platforms have treated people as products—optimizing for engagement over wellbeing, prioritizing profit over safety, and concentrating power in the hands of a few companies and individuals rather than sharing it with users and communities. We are now paying for these decisions as the attention economy pervades society and rewards performance over substance.
Bluesky's mission is to fix the structural issues with social media by transitioning from platforms to protocols. We believe the future of social networking should be open, give users genuine control over their data and experience, and create accountability through transparency. That means building differently—not just better moderation policies, but a fundamentally different architecture that puts power back in users' hands.
In 2025, as Bluesky grew from 25 million to 41 million users, we improved the trust and safety infrastructure to better enable that mission. Here's what that looks like in plain language:
Our moderation approach is human-centered. We use automated systems to catch the bad stuff at scale—spam networks, child exploitation, and coordinated attacks. For everything requiring context, nuance, or cultural understanding, trained human moderators make the call. When you report harassment, hate speech, or other complex harms, a person reviews it against our Community Guidelines.
We are building transparency into our systems. Our new strike system is continually evolving to inform you which policy you violated, how serious it was, and where you stand. When we make mistakes—which we will—our appeals process involves human review, not just automated rejection.
The most important developments in 2025 were:
We dramatically reduced toxic content through proactive design. By October, our reply filtering system had decreased user reports of anti-social behavior by 79%. This wasn't about removing speech or banning users—it was about using product design to reduce the visibility of toxic replies while keeping them accessible to those who want to see them. Small design choices, thoughtfully implemented, created measurably healthier conversations.
We implemented age assurance across multiple jurisdictions. This was genuinely hard. Nearly every country and US state has different, often conflicting requirements, and some well-intentioned laws also created serious barriers to privacy and free expression. We built custom paths for the UK, Australia, and multiple US states, threading the needle between legal compliance and our values. When Mississippi's law forced impossible choices, we initially blocked access rather than compromise on privacy—then built better systems and restored access for adults later in the year.
We scaled our moderation operations to match our unprecedented growth. In 2025, users created 1.41 billion posts—61% of all posts ever made on Bluesky. We maintained 24/7 moderation coverage, reviewed 9.97 million user reports, applied 16.49 million labels, and removed 2.45 million pieces of violating content. We brought our spam and bot detection capabilities to a level where misleading content reports dropped month over month even as our user base grew. We built a verification system that includes both centralized verification from Bluesky and a network of Trusted Verifiers—news organizations, universities, and cities have signed up institutions that can verify their own communities.
This work required substantial investment and difficult choices. We were forced to make hard trade-offs between speed and accuracy, between removing harm and preserving expression. We didn't get everything right. But we further built the foundation for trust and safety systems that can scale while staying true to our mission.
Thank you for being part of building something better.