Research Hub
The Research Hub is a public, freely readable library of citation-grounded briefs the agent has already produced — organized into three registers, and designed so any brief can become the seed of your own live research run.
#A public library of briefs
Every brief in the Hub is a finished piece of the agent's work: a research question the agent has planned, searched, read in full text, weighed, and written up with numbered, verified citations. Unlike a run in your workspace, these briefs are published for anyone to read — no account required — so the Hub doubles as a browsable showcase of what an evidence-grounded answer looks like and as a starting point for your own questions.
- Free and open. Briefs are readable by the public; each one is marked as freely accessible.
- Citation-grounded. Claims trace back to the literature the agent read, the same way they do inside the app.
- Searchable and dated. Filter within a register by keyword, and see when each brief was published.
- Continuable. Any brief can open a fresh live run seeded with its question, so the library is a launchpad rather than a dead archive.
Because the briefs are static, published documents, they reflect the evidence available when the agent wrote them. To bring a topic up to date, or to push on it in a direction the brief did not take, continue it as a live run (see below).
#The three registers
The Hub is divided into three registers — Hypotheses, Methods, and Pathways — each answering a different kind of research question. A register is a lens on the same underlying agentic work, not a different product; the difference is the shape of the question and the output the agent is asked to emphasize.
#Hypotheses
Evidence-grounded, testable hypotheses on emerging biomedical topics. Each brief lays out a claim the literature makes plausible, the evidence behind it, and where that evidence is thin. Open one in the app to challenge the hypothesis, expand its evidence base, or adapt it for your own grants and protocols.
#Methods
Advanced experimental strategies and analytical frameworks. Each brief surveys how a technique works, when it applies, and how it compares to alternatives. Continue one in the app to drill into a specific protocol, compare techniques head to head, or adapt a methodology to your own dataset.
#Pathways
Curated insights into disease-relevant molecular pathways and gene networks. Each brief traces mechanisms and the targets along them. Continue one in the app to follow a mechanism further, branch into a specific target, or compare a pathway across conditions.
| Register | Answers | Typical next step in the app |
|---|---|---|
| Hypotheses | What is a testable idea worth pursuing here? | Challenge it, or adapt it for a grant or protocol. |
| Methods | How would I actually do this experiment? | Compare techniques, or fit one to your dataset. |
| Pathways | What mechanism and targets are involved? | Trace a mechanism, or compare across conditions. |
#Continuing a brief as a live run
The Hub is not read-only. Every brief carries a Continue in the app action that turns a published document into a new, live agentic run of your own. This is the bridge from “something the agent already answered” to “a question the agent is answering for you, right now.”
Here is what happens when you continue a brief:
- A new run is created. Continuing does not edit the published brief or reopen the original run — it starts a fresh live run in your workspace.
- The composer is prefilled. The app seeds the message box with the brief's research question, so you are not starting from a blank page.
- You review and send. Nothing runs automatically. You can edit or refine the prefilled question first, then send it to launch the agent.
- The agent works from scratch. Once sent, the run plans, searches, reads full text, weighs evidence, and verifies citations against the current literature — the same workflow as any other run.
#How the Hub relates to your own runs
A published Hub brief and a live run in your workspace are the same kind of artifact at two stages of life. The brief is a snapshot the agent has already produced and made public; a live run is the agent working a question in real time for you. Continuing a brief simply moves it from the first state to the second.
- Reading a brief is passive and public — no account, no compute, evidence frozen at publication time.
- Continuing a brief is active and yours — a fresh run against the current literature, fully steerable before it starts.
In short: browse the Hub to see what the agent can do and to find a well-framed starting question, then continue any brief to make it your own.