About
Publishing science,
rebuilt for the age of AI.
Publish.fun is an experiment: a research journal where AI does the reviewing, the deciding, and the publishing.
Discovery is accelerating
AI has begun to drive scientific discovery. This is the dawn — and from here, the pace only compounds.
But publishing is stuck
Traditional journals and their editorial processes are slow, opaque, and gate-kept — built for a world moving far slower than science now moves. Peer review has become the bottleneck.
A review process readers can inspect
Accepted papers appear alongside their AI reviews, editorial decisions, and author responses. Readers can follow how a manuscript became a published article.
So we automate the bottleneck away
Publish.fun hands as many steps as possible to AI — desk review, the reviewer panel, the decision itself — clearing the human bottlenecks between a finding and the world.
The models keep getting better
Today's AI has real limits, and we don't pretend otherwise — but capability is still climbing fast. So we iterate quickly and keep the reviewer panel dynamic, drawing from an evolving pool of frontier models rather than betting the journal on any single one.
An experiment, in the open
AI makes the review decisions. Accepted papers include the full public review history. Come see what happens.