AI Open-Access Journal

An audited AI pipeline for scientific publishing. Researchers pre-register study designs, submit results, and the platform generates papers using Claude — with citation verification, open peer review, and a full audit trail.

Live: arxai.science

How it works

1. Pre-register a study design (empirical, simulation, replication, or negative results)
2. Submit results as data artifacts with a manifest
3. AI generates the paper section-by-section from the pre-registration and results, enforcing faithful reporting
4. Citations are verified against Semantic Scholar and CrossRef
5. Open peer review with attributed reviews and structured scoring
6. Publication with full transparency: pre-registration, data, audit trail, and reviews all visible

Why

Scientific publishing has a principal-agent problem: the people who run studies are the same people whose careers depend on results looking good. AI has no career — it won't bury a negative result. The audit trail makes the entire pipeline inspectable, replacing trust-by-proxy with verifiable transparency.

Tech stack

Next.js 15 / React 19 / TypeScript / Tailwind CSS / Drizzle ORM / Neon Postgres / NextAuth.js / Claude Opus 4.6 / Vercel

Development

npm install
npm run dev

Requires environment variables — see .env.example.