Agents For Academia

Hello world 👋 Agents4Academia is a community‑led effort to explore and build open‑source agents for academic and research work — by researchers, for researchers. agents4academia.org

@klarakaleb @harit_v @yeewhye Github: github.com/Agents4Academi… Website: agents4academia.org/events/2026-ox… Interest form: forms.gle/iausY6K72XGWAy…

This is episode 1⃣. Contribute to a project, propose a new one, or run an Agents4Academia event where you are. Get in touch: @klarakaleb @harit_v @yeewhye

First up a hackathon: 20+ researchers · two weeks · @OxfordStats × @NUSComputing × @NTUsg and tokens on tap by @AnthropicAI . Outcome: Five open‑source AI agents solving common workflow frustrations across the whole research lifecycle. 🧵

Hello world 👋 Agents4Academia is a community‑led effort to explore and build open‑source agents for academic and research work — by researchers, for researchers. agents4academia.org

To see it in action, we ran Prior on our own field — agents for the scientific process. 152 papers · 581 contributions · 989 typed relations · 9 emergent communities (peer review, benchmarks, safety, and more). The (rough) structure of the field emerged right in front of us.

Then, a graph, not a summary: 2⃣ Contributor — full text → atomic claims with supporting quotes. 3⃣ Cartographer — connects claims across papers into a typed graph (supports · extends · contradicts). 4⃣ Navigator — cited, grounded answers.

1⃣Scoper — open, queryable scholarly data is getting rich enough (thanks to @OpenAlex_org & @SemanticScholar). Prior's first subagent mimics what we do manually — snowball from anchor papers, follow citation trails, chase references. Now automated + structured.

Day 1⃣ of our @agents4academia hackathon, we asked the participants: what are the most frustrating parts of your workflow? Almost every single one: keeping up with the literature. The firehose is impossible. The trust is breaking (👋AI slop). So we built something. Meet Prior 🧵