Bluesky Launches Attie, an AI-Powered Custom Feed Builder Using Anthropic's Claude
Key Takeaways
- ▸Attie is a standalone AI assistant app—not integrated into Bluesky proper—that lets users build custom feeds through natural language prompts powered by Claude
- ▸The app leverages Bluesky's open AT Protocol ecosystem to access shared user data, enabling personalized recommendations without requiring users to write code
- ▸Future roadmap includes the ability for users to 'vibe-code' their own social apps and build tools for other developers on the atproto network
Summary
Bluesky has unveiled Attie, a standalone AI application that enables users to create custom social media feeds and eventually build their own apps using natural language commands. Powered by Anthropic's Claude, Attie launched in beta at the Atmosphere conference and integrates with Bluesky's underlying AT Protocol (atproto), allowing users to leverage open ecosystem data to personalize their social experience without requiring coding knowledge. The app represents Bluesky's expansion beyond its core social network into developer tools and represents a strategic pivot for former CEO Jay Graber, who stepped down to lead the innovation team building new products on the platform. Interim CEO Toni Schneider emphasized that Attie is designed as a "people-focused" AI product that democratizes feed curation and app development for the broader atproto ecosystem.
- Jay Graber, Bluesky's former CEO, now leads the innovation team building Attie and new products, having stepped back from CEO duties to focus on product development
Editorial Opinion
Attie signals an intriguing shift in how social networks might evolve under decentralized protocols—empowering users to shape their own feeds through conversational AI rather than algorithmic black boxes. By combining Anthropic's Claude with atproto's open data architecture, Bluesky is positioning itself not just as a Twitter alternative, but as a platform for broader creative experimentation. The move underscores a broader trend of AI-assisted customization replacing one-size-fits-all algorithms, though success will hinge on whether non-technical users actually adopt such tools and whether the open ecosystem can maintain quality and safety at scale.
