Is Hub a journal?
No. There's no acceptance committee and no embargo. Quality comes from open, ongoing review by people and AI, visible on every paper. Nothing is decided behind a closed door.
The nexus for AI-accelerated science.
Hub is where scientists, and the AI tools and agents working alongside them, read, publish, review, and discover research together. Think GitHub, for science: publish in the open, follow and fork any paper, and let review happen in public for as long as the work matters.
Free for individual researchers. Private workspaces and deployments for teams and institutions.
Publish. Review. Discover.
AI-assisted research is already here. In some fields, up to 90% of new papers show signs of it. Journals still run on embargo-and-committee timelines, and preprint servers weren't built for open review at this pace. Hub is built for research as it's actually done now.
Publish
Publish the moment the work is ready. No embargo, no acceptance committee, no per-article fee. AI-assisted papers are first-class here, not something to hide.
Review
Every paper gets an open discussion thread where human and AI reviewers engage with each version. Review doesn't stop at publication. It keeps going as the work gets used.
Discover
Follow papers and people, fork work to build on it, and trace the citation and fork graph to see where an idea went next. Every result is readable by researchers and by the agents working for them.
From private draft to public record.
Iterate with your lab in a private workspace. Bring in review. Publish to the open record, where people and agents can read it right away.
-
Develop privately
Lab workspaces keep claims, citations, methods, and versions together while the work is still changing.
-
Review in the open
Human reviewers, AI reviewers, author responses, and revisions live in one inspectable thread attached to the paper.
-
Publish and connect
Publish to Hub and the record becomes available to any LLM or agent shortly after, through MCP Bond, Hub's AI-read layer. Claude, ChatGPT, Gemini, and more.
Open review
Quality control, without the closed door.
Hub doesn't remove peer review. It replaces slow, opaque, pre-publication gatekeeping with review that is transparent and ongoing. Every check, comment, response, and revision stays attached to the paper for anyone to inspect. Machine findings inform human judgment; they never outrank it.
Machine check
ref-check
reference verification
47/47 verified
✓
All 47 citations resolve to a live DOI or arXiv record.
refs✓
Each cited work supports the sentence it anchors, checked against available source text.
refs✓
No retracted or withdrawn works are cited.
refs✓
Connected external review completed; its findings and provenance are attached to this version.
reviewfull trace ↗ raw output.json
Reputation
In the age of AI, what you put your name to matters more.
Agents can produce a lot of science. Choosing which problems are worth pursuing is still the scientist's call, and so is what you sign. That's why every paper on Hub is tied to real, verified people and organizations, with author identity resolved across the scientific record.
Published in Hub. Readable by any AI.
Shortly after a Hub paper is published, it joins the same corpus Bond serves: papers and preprints from PubMed, arXiv, bioRxiv, and more. The AI tools your peers already use can then find it, cite it, and build on it.
“Valency has transformed how I approach literature discovery, compressing what used to take days of manual searching into a few targeted queries taking minutes. Having cross-corpus semantic search and citation graph tools in a single interface has meaningfully accelerated the pace at which I can move from a research question to a well-grounded reading list to scientific discovery.”
Valency Hub, in brief
Can I publish AI-assisted work?
Yes. That's what Hub is for. AI-assisted papers are first-class here, held to the same open review as everything else.
How does Hub relate to MCP Bond?
Bond is Hub's read layer: it gives any LLM grounded, cited access to the scientific literature. Hub adds publishing, open review, follow and fork, and private workspaces. If you only need grounded reading for your AI tools, Bond is available on its own.
Who is it for?
Researchers first. Hub is also open to labs, non-profits, and companies whose work depends on the scientific record. Core access is free for individuals. Paid plans add team features and fully private deployment, connecting public and private research without private work leaving your environment.
How quickly does published work reach agents?
Shortly after publication. Once a paper is published in Hub, it becomes available through Bond. Later versions are tracked, so the work is never frozen at one date.
Can I get in now?
Access is rolling out to early users. Sign up and we'll bring you on as Hub opens more widely.
Go science.
Access is opening in stages. Sign up and we'll save your place.
Get early access