Ronen Tamari, Cosmik

We’ll present Semble, a kind of “Spotify for research” enabling researchers to curate shareable collections, create knowledge trails that others can explore and extend, and discover relevant work shared with the network. Built on ATProto, Semble offers researchers data portability and an open API designed for extension. We'll discuss how Semble enables new kinds of research tooling, from living semantic citation graphs to collaborative review and annotation. We'll also share how Semble connects to Cosmik's broader work on collective intelligence, and opportunities for collaboration across the ATProto science ecosystem.

An atproto.science workshop

Asst. Prof. Maria Antoniak (Uni. Colorado Boulder)

We will present Lea, a social app for researchers built on ATProto. Lea has many custom features for researchers, including

1. paper tracking, discovery, and discussion pages,

2. customized and verifiable profiles for researchers,

3. extensive safety and moderation features to keep discussions calm and productive.

We'll discuss our goals, challenges, and open questions for Lea.

An atproto.science workshop

Assoc. Prof. Aaron White (Uni. Rochester), Chive

Chive is a decentralized preprint service featuring threaded review, formal endorsements, and a community-curated field taxonomy—all as portable ATProto records users own. Chive provides a rich plugin interface, making it imminently extensible as the ATmosphere grows. It currently provides builtin plugins for integration with existing ATProto services, such as Semble, Leaflet, and WhiteWind.

Chive will be launching a call for alpha testers in the next week or so, with a plan to roll out an open beta around the time of the workshop.

An atproto.science workshop

Dr. Sean Jungbluth, BioKEA

BioKEA will demo a proof-of-concept automated reviewer for scientific papers.

Provided with an arXiv ID or PDF, it extracts structured text, pulls context from Semantic Scholar and PubMed, and outputs a critical review flagging errors, missing context, and citation gaps.

We'll run it live, preview some features, and discuss publishing review artifacts on ATProto.

An atproto.science workshop

Travis Simpson, Skysquare Inc

Skysquare exists to make sense of the digital public square via Bluesky. Launching this summer as a Chrome extension, Skysquare overlays social media conversation onto the webpages people are discussing, allowing readers to experience our online conversation in richer context.

When a page has been shared or talked about on Bluesky, Skysquare associates that discourse with the page itself and reveals the surrounding conversation directly in the browser.

Instead of chasing links through feeds, readers and writers can immediately see who is talking about what in relation to the material they are reading, turning the web into a socially annotated layer of context.

This talk will demonstrate the extension in action and explore how tools like Skysquare can help researchers, scientists, and educators connect conversations directly to the sources being discussed.

An atproto.science workshop

Alex Garcia-Joyner

One of the biggest problems in society today is division driven by social / societal ills, information overload, and platforms that amplify echo chambers.

In this talk we'll discuss how we're addressing all these problems at ViewSift through our new atproto social-research platform and teaching practical techniques to have healthy discussions around controversial topics that heal division.

An atproto.science workshop

Anish Lakhwara

Seams.so demo and live workshop

An atproto.science workshop

Dr. Emily Hunt

What's the point of doing science if you can't tell anyone about it? I'll present The Astrosky Ecosystem, a community project by astronomers to democratize social media access for the space science & space fan communities. I'll talk about our 30 months of running custom feeds, as well as our future plans to start PDS hosting and even venture towards an astrophotography appview.

An atproto.science workshop

Dr. Teon Brooks

We will discuss new advances in computing education. We will highlight projects that teach computing and data science at scale using Pyodide. We propose a new infrastructure model that builds on these advances and incorporates the ATmosphere as an identity model and content storage. Using ATmosphere services like Tangled and Blacksky, we can truly democratize computing education.

An atproto.science workshop

Dr. Scott McGrath

AMIA's vibrant Twitter backchannel fragmented post-X. This talk details our stalled effort to migrate the community to Bluesky. Despite a guide and conference launch, the "cold start" problem hindered adoption. I’ll share lessons learned, discuss migration barriers, and outline revised strategies to rebuild our clinical research network.

An atproto.science workshop

Jay Patel

Research synthesis, a desirable culmination of primary research, is notoriously slow, error-prone, and disconnected from the network of potential contributors.

Now, ATProto offers a digital foundation upon which to recruit collaborators (Bluesky feeds), assign micro-tasks (discourse graphs), author reports (@leaflet.pub) , and acknowledge contributions.

An atproto.science workshop

Sophie Greenwood (Cornell Tech), Skygest

The AT Proto ecosystem empowers novel social media research. Our research showcases three promising directions in this space: 1.

1. Experiments on self-hosted feeds (Paper Skygest, @paper-feed.bsky.social, findings and customization interface)

2. Experiments on existing feeds via collaboration with feed designers (collaboration with @graze.social and @aendra.com),

3. Observational analyses of social media (SAEs on AT Proto posts)

An atproto.science workshop

Francisco Carvalho

The Community Archive is a community-owned dataset of contributed Twitter data used to study how ideas spread in online communities.

We developed methods to extract "narrative strands" — coherent lines of discourse where ideas evolve and build on each other over time. These techniques generalize directly to atproto datasets and could help Bluesky communities understand their own emerging canons.

An atproto.science workshop

Billy Pierce

Presenting a working paper on the functional (de)centralization of Bluesky.

This includes the extent to which AT Protocol infrastructure is owned by entities other than Bluesky Social (through mapping ownership of PDS endpoints) and the implications of this (de)centralization on Bluesky Social's goals. Following the presentation, there will be a Q&A and discussion!

An atproto.science workshop

Research institutions face a structural gap between institutional knowledge (formal, slow, siloed) and the personal knowledge networks (informal, fast, distributed) of their faculty and students. The Atmosphere is the first infrastructure that can bridge this gap — not by centralising personal knowledge, but by making both layers speak the same protocol.

Part of Building ATScience (Lounge, 14:45-15:50).

(A repost of v4 of workshop proposal published on my wiki, where the latest version can always be found.).

At the top of my mind is the recognition that the Atmosphere's killer app is not Bluesky - it's interoperability. But it's not enough to have some of your staff using interoperable Atmosphere apps - it requires coordination and knowledge management techniques to get the most out of it.

So in this session we'll try to pick up ideas from all the sessions preceding it, and weave together some ideas and practical steps for research organisations seeking to get the most out of the Atmosphere.

I'd really love your ideas before the event so I can give you the floor on the day, so please post a comment here and/or select some text, click "Share via Bluesky" and post your comment or question to Bluesky, adding @mathewlowry.eurosky.social to ensure I spot it ;)

Opportunities

The Atmosphere provides new opportunities for all organisations to better integrate institutional communications, knowledge management and sensemaking with the personal communications, knowledge management and sensemaking of their staff (past, present and future). For educational institutions, moreover, their students may offer another layer of opportunity.

Getting this right could allow your institution to better tap the knowledge and networks of your staff and students, reinforcing research communities and knowledge transfers. There have been several technological solutions to exploiting this opportunity, from centralised knowledgebases to linked data. I want to explore whether the Atmosphere provides better answers.

The key differences are:

• The institution doesn't have to own knowledge to benefit from it
• Staff/students don't have to donate anything, or even make extra effort to contribute knowledge - it flows frictionlessly from their personal data servers (PDSs) if they choose
• When someone leaves your organisation, connections don't have to break.

Challenges & Solutions

Exploiting this opportunity fully, however, means interlinking infrastructure (HR systems, website content management and other IT infrastructure) and encouraging/training staff in the use of a variety of Atmosphere apps, from Bluesky to Semble, Margin to standard.site.

I think, however, that this makes it sound harder than it is. Moreover, you don't have to do everything at once - like any programme, one should start small and build over time, learning as you go.

Custom feeds & Starter packs

Last November in Berlin for example, I explored what I now see as many organisations' first real step into the Atmosphere: using Bluesky starter packs and custom feeds, an organisation can both create a powerful Bluesky presence and benefit from the trust in their faculty's personal networks to convene highly valuable conversations arounds topics of importance to the institution.

Building on the personal networks of faculty, this can amplify the work and reputation of the institution and its faculty, and help the institution discover new knowledge and build better networks.

More: How newsrooms, scientific institutions & governments can best use Bluesky (November 2025)

Website as community scaffolding

The above strategy, however, is building entirely within Bluesky. Benefiting from other apps requires using your institutional website as scaffolding to integrate your community of staff and students, the apps they use, and the knowledge and networks flowing through them.

Your website, in other words, should be much more than just an online brochure - it should become a node in the Atmosphere, and a "homebase" for your institution's community and the knowledge they're sharing and discussing.

When I first started exploring this, I discovered that farming out your website's interactivity to the Atmosphere brings you greater reach and engagement, while radically simplifying your website publishing technology (Simple sites, powerful communities). I suspect this is just the tip of an iceberg.

Collective knowledge discovery

Note that that pattern only involved standard.site and Bluesky, and there's much more to the Atmosphere than that.

I'm wondering, for example, how an organisation could aggregate its community's use of Semble, Margin, Skysquare and Seams (all being presented in Vancouver) to tap into the community's collective sensemaking and knowledge discovery, and then add value to it somehow.

Similarly, tools like Sill allow individual users to discover knowledge popular with each user's social graph. What would happen if you aggregated this knowledge together in a form of collective listening?

Knowledge publishing

Unsurprisingly there's a lot going on in the Atmosphere for scientific publishing.

An early workshop, for example, will include a presentation of Chive: "a decentralized preprint service featuring threaded review, formal endorsements, and a community-curated field taxonomy - all as portable ATProto records users own".

With Chive, each researcher's record travels with them as they move between institutions, rather than being locked into a single journal or repository. But I think it's fair to say that the institution which hosted that research would like it to appear on their website as well.

Moreover, from the sounds of it a "community-curated field taxonomy" sounds like distributed sensemaking, echoing the collective listening made possible by Sill, but applied to scholarly categorisation rather than news and links.

Similarly, Morning Session 2 will see:

• "Automated science coordination with ATProto", which will present ATProto as "a core component of the coordination.network stack... [enabling] direct nano-publishing from the lab bench" -
• "atdata: Distributed datasets over atproto", which will explore "how atproto enables a step-change for the use of large-scale, distributed, open, interoperable scientific datasets"

What I love about all of these apps is that while the data on them is owned by the researchers who put it there, it can not only be used on the app itself. The content's on the users' PDSs, and so can also pulled into other apps, where it can be combined with other knowledge from other apps to add further value.

This is where the idea of University website as scaffolding might become useful: with a University's various researchers creating content on a variety of different apps, it might be helpful to bring it all together in a single website, while respecting the researcher's ownership of the data.

Suggestions, Questions?

As mentioned before, please help me flesh this out before the event - I would be very happy to give you a slot in this workshop.

Either post a comment here and/or select some text, click "Share via Bluesky" and post your comment or question to Bluesky, adding @mathewlowry.eurosky.social to ensure I spot it.