Bluesky has launched Attie, an AI-powered app that lets users design custom social feeds without writing code, marking the first major product from the company's new innovation division. Built on Anthropic's Claude and running atop Bluesky's open AT Protocol, Attie represents a deliberate pivot: away from competing with Meta and X on algorithmic engagement, toward a model where users themselves become feed architects. The move lands as Bluesky crosses 43.4 million users and secures $100 million in additional funding—enough runway, executives claim, to build infrastructure that existing platforms have no incentive to create.
Dispatch
San Francisco, March 28, 2026 — TechCrunch's Sarah Perez reported from the Atmosphere conference, where Bluesky's former CEO Jay Graber (now chief innovation officer) and CTO Paul Frazee unveiled Attie for the first time. The reporting establishes both the technical architecture and the philosophical intent behind the product.
The team from Bluesky has built another app — and this time, it's not a social network, but an AI assistant that allows you to design your own algorithm, create custom feeds, and, one day, vibe-code your own app. [1]

The mechanism is straightforward: users sign in with their atproto credentials (the same login that works across Bluesky and other apps built on the protocol), then chat with Claude in natural language to specify what they want to see.
With Attie, anyone will be able to build their own custom feed just by typing in commands in natural language, the same as if they're chatting with any other AI chatbot. You can ask Attie questions, like what posts you might like to see or repost, and you can use the app to curate your own custom feed, personalized to you. [1]
Interim CEO Toni Schneider (who is also a partner at True Ventures, a Bluesky backer) emphasizes that this is not a feature buried inside Bluesky—it is a standalone product, the first from Graber's newly formed team. Schneider frames the offering as user empowerment rather than corporate convenience:
You control it, you shape it, without having to write code or know how to set up these feeds. It's the beginning of just having a lot more people be able to build on top of the Atmosphere. [1]
Graber's own framing is sharper—a direct critique of how incumbents weaponize AI:
We think AI should serve people, not platforms. An open protocol puts this power directly in users' hands. You can use it to build your own feeds, create software that works the way you want it to, and find signal in the noise. [1]
No major outlet has yet offered a contrasting account. This analysis draws from the single source above, supplemented by context from Bluesky's disclosed funding and user metrics.
What's Really Happening

The Real Stakes
For Bluesky's survival: The company has 43.4 million users but no proven monetization engine [1]. The $100 million in funding provides "three-plus years of runway," according to Schneider [1], but that clock is ticking. If Attie can drive engagement and demonstrate that users will pay for control (or that hosting services generate revenue), Bluesky moves from funded-but-fragile to sustainable. If Attie becomes a niche tool used by a few thousand power users, it does not solve the core problem.
For the protocol economy thesis: Attie is an experiment in whether ordinary users—not just developers—will adopt decentralized infrastructure if given the right UI. The AT Protocol has been live for years; Bluesky's own custom feeds feature existed before Attie. What Attie tests is whether AI-assisted curation is the missing link—the thing that makes protocol-based social networks feel as intuitive as centralized platforms. Confirmed: Bluesky's team believes this. Projected: If Attie gains 100,000+ active users within six months, other protocol-based networks will copy the model within a year.
For the AI-in-social-media narrative: This is a rare case of AI serving user interests rather than platform interests. Meta, TikTok, and X use AI to maximize engagement and data extraction. Bluesky is using AI to maximize user choice. One scenario: This becomes a genuine differentiator, attracting users fatigued by algorithmic manipulation. Another scenario: Users find feed-building tedious and revert to curated feeds anyway, at which point Bluesky's differentiation evaporates.
For crypto investors: Schneider explicitly denies any crypto integration despite multiple crypto investors backing Bluesky [1]. This is notable because it signals that the company is not chasing the "blockchain social network" narrative—it is building genuine infrastructure, which happens to appeal to crypto investors who value decentralization. This distinction protects Bluesky from the crypto-scam associations that plague other decentralized platforms.
Industry Context
The WordPress parallel is both apt and misleading. Schneider notes that WordPress generates over $10 billion annually despite being fully decentralized [1]. True—but WordPress succeeded because it solved a real, urgent problem (self-hosted publishing) that centralized platforms (Blogger, Typepad) were not solving well. The question for AT Protocol is whether "user-controlled feeds" is an urgent problem for 43.4 million people, or a niche preference among power users and privacy advocates.
Custom feeds already exist on Bluesky. The company has offered custom feed functionality for months. What Attie adds is the AI layer—the ability to describe what you want in English rather than writing code or clicking through UI. This is a meaningful UX improvement, but it is not a new capability. The bet is that the UX improvement is large enough to move custom feeds from 5% adoption to 30%+ adoption. No data on current custom feed usage has been disclosed.
Monetization pressure is real. Bluesky cannot remain a funded-but-unprofitable platform indefinitely. The company has signaled three possible routes: Attie subscriptions, protocol subscriptions for power users, or hosting services for communities. None of these has been tested at scale. The WordPress analogy suggests that ecosystem monetization (third-party developers and services) could eventually dominate, but that requires a much larger developer base than Bluesky currently has.

Impact Radar
Watch For
1. Attie adoption metrics within 90 days. Bluesky has not disclosed how many conference attendees became beta testers or how many have used Attie since launch. If the company announces 50,000+ active users by June 2026, the product is on track. Below 10,000, it is a niche tool [1].
2. Pricing announcement. Schneider stated that no decision has been made on whether Attie will charge users [1]. If the company announces a paid tier within six months, it signals confidence in retention and monetization. If Attie remains free indefinitely, it is a user acquisition tool, not a revenue engine.
3. Third-party apps adopting Attie-like features. The AT Protocol is open; any developer can build a Claude-powered feed curator. If Bluesky's competitors on atproto (or ActivityPub-based platforms like Mastodon) launch similar products within 12 months, it will confirm that this is a viable pattern. If no one else copies it, it suggests limited market demand [1].
4. Bluesky's monthly active user growth rate. Attie is designed to increase engagement and retention. The company should disclose MAU figures quarterly. If growth accelerates post-Attie launch, the product is working. If growth stalls or slows, Attie has not solved Bluesky's core retention challenge.
Bottom Line
Attie is not a product that wins markets. It is a product that tests a hypothesis: that users will prefer agency over algorithmic convenience. If that hypothesis holds, Bluesky has a path to differentiation and eventual profitability. If it does not—if most users find feed-building tedious and revert to curated feeds—Bluesky remains a well-funded but strategically stuck social network with no clear moat against incumbents. The next six months of adoption data will be far more revealing than Bluesky's funding round or user count.