Bluesky Unveils Attie: AI‑Powered No‑Code Feed Builder

Why Attie Matters in the Decentralised Social Landscape

Bluesky has spent three years championing an open, federated social‑media protocol that removes the single‑gatekeeper model of traditional platforms. The UBOS platform overview shows how open standards empower developers and users alike. Attie extends this philosophy by giving every user a conversational AI that can shape their timeline without writing a single line of code.

Most mainstream networks rely on opaque recommendation engines that optimise for watch‑time, not user intent. Attie’s “people‑first” approach directly counters that trend. As former CEO‑turned‑Chief Innovation Officer Jay Graber explained at Atmosphere, “We think AI should serve people, not platforms.” This ethos resonates with tech‑savvy social‑media enthusiasts who crave transparency and control.

With Attie, you can simply type:

• “Show me posts about sustainable tech from the last week.”
• “Create a feed that mixes indie‑music news with AI research updates.”
• “Hide anything that mentions crypto scams.”

The AI interprets these prompts, writes the underlying AT Protocol query, and instantly surfaces a custom timeline—no coding required.

How Attie Works: From Login to Live Feed

1. Single‑Sign‑On via Atmosphere

Users log in with their existing atproto credentials—the same identity they use for the Bluesky app. This unified login lets Attie instantly read a user’s interaction history, follows, and preferences across the entire ecosystem.

2. Claude‑Powered Understanding

Attie integrates the OpenAI ChatGPT integration philosophy but swaps in Anthropic’s Claude model, known for its safety‑first training. Claude parses natural‑language requests, maps them to AT Protocol query syntax, and returns a feed definition.

3. Instant Feed Generation

The generated query runs against the decentralized network, pulling in posts that match the criteria. Users can preview, tweak, or save the feed for later use.

4. Future “Vibe‑Code” Layer

Later releases will let power users export their feed definitions as reusable “vibe‑code” modules—mini‑apps that can be embedded in any atproto‑compatible client. Think of it as a plug‑and‑play widget marketplace, similar to WordPress plugins.

The Team Driving Attie’s Launch

Key figures include:

• Jay Graber – Former Bluesky CEO, now Chief Innovation Officer, leading product vision.
• Paul Frazee – CTO, architect of the AT Protocol, overseeing Claude integration.
• Toni Schneider – Interim CEO and partner at True Ventures, providing strategic guidance and linking the project to Bluesky’s funding roadmap.

According to Schneider, the Attie team began development only a few months ago, after Graber stepped back from day‑to‑day operations to focus on product innovation. “She realized she wanted to build, not just run the company,” Schneider told TechCrunch.

For readers interested in how AI agents can be embedded in other workflows, check out the ChatGPT and Telegram integration on UBOS.

Funding, Roadmap, and Monetisation Plans

Bluesky’s recent $100 M funding round—closed in late 2025—gives the protocol a three‑plus‑year runway. The capital will be allocated to three core pillars:

1. Privacy Enhancements – User‑controlled data sharing and selective feed visibility.
2. Scalable Infrastructure – Supporting the growing atproto network, now 43.4 M users.
3. Monetisation Experiments – Subscription tiers for premium feed‑building tools, hosting services for private communities, and possibly a marketplace for vibe‑code modules.

Schneider emphasises that crypto integration is off the table, despite the presence of crypto‑focused investors. “Our backers love decentralisation, not token economics,” he notes.

For a glimpse of how UBOS structures its own pricing, see the UBOS pricing plans. The modular approach mirrors what Bluesky hopes to achieve with Attie.

Impact on Users and the Broader AT Protocol Ecosystem

For End‑Users

Attie hands control of the algorithm back to the individual. No more “black‑box” timelines; users dictate exactly what they see. By specifying filters in plain English, they can sidestep spam, misinformation, and unwanted topics.

For Developers

Attie’s natural‑language to query translation is built on open standards. Developers can inspect, fork, or improve the underlying code. The upcoming vibe‑code feature could spawn a marketplace of plug‑and‑play social widgets, akin to the UBOS templates for quick start.

UBOS already showcases a rich template ecosystem; for example, the AI SEO Analyzer lets marketers automate keyword research, while the AI Article Copywriter generates SEO‑friendly copy in seconds. Attie could become the social‑media counterpart of these tools.

Attie’s Role in the Growing Decentralised Social Movement

Projects like Mastodon’s “timeline‑filters”, Lens Protocol’s “smart‑feeds”, and ActivityPub‑based AI assistants are all experimenting with user‑controlled timelines. Bluesky’s advantage lies in its single, unified protocol (AT Protocol) that already powers a sizable user base, making Attie the first AI‑first feed builder that can operate at scale without fragmentation.

Schneider likens the Atmosphere ecosystem to WordPress: “At the centre of the Atmosphere is a completely open system, so anybody can participate.” He predicts a $10 B annual ecosystem similar to WordPress’s plugin market.

For businesses looking to harness AI‑driven marketing, the AI marketing agents on UBOS illustrate how AI can be applied beyond content creation—into campaign optimisation, audience segmentation, and real‑time analytics.

What to Expect Next from Attie

• Beta Expansion – Private beta will roll out to more Atmosphere attendees over the next two months, with a public beta slated for late 2026.
• Feature Additions – Sentiment‑based filters, collaborative feed building, and analytics dashboards for power users.
• Monetisation Trials – Potential subscription tiers for advanced feed‑creation tools and hosting services for private communities.
• Vibe‑Code Marketplace – A future app store where developers can sell or share custom feed modules.

Early adopters can already experiment with the Talk with Claude AI app on UBOS, which demonstrates Claude’s conversational capabilities in a different context.