Journal in a Box (JiaB) is free, open-source software for running a complete peer-reviewed scholarly journal: submission, review, editorial decisions, publication, and post-publication evaluation.
Cheap
Journal in a Box is free software. It runs on server infrastructure that costs less than $5 a month. The optional AI-assisted features use pay-per-use pricing from well-known LLM providers and remain negligible for typical journal volumes.
Deploy in minutes
JiaB ships as a single self-contained binary. There is no database server to install, no application server to configure, no dependencies, and no external runtime to maintain. Drop the binary on any Linux server, run jiab init, and your journal is live.
Automated editorial checks (optional)
Every submission can be screened by a suite of configurable checks before it reaches the editorial queue:
- Anonymity — flags author names, affiliations, acknowledgements, and identifying self-citations that would compromise blind review.
- Pre-registration — detects mentions of pre-analysis plans or registrations (OSF, AsPredicted, ClinicalTrials.gov) and requires the registration document to be uploaded.
- Replication materials — identifies manuscripts with original statistical analysis and requires a public replication URL.
- PAP comparison — when a pre-analysis plan is uploaded, an LLM-assisted comparison across configurable dimensions highlights deviations between the registered protocol and the submitted paper.
- Manuscript quality - LLMs can be used to flag potential issues in the scientific content of manuscripts. This feature is similar to refine.ink, coarse.ink, and Reviewer 2.
Checks are implemented as plugins. Journals can disable built-ins, adjust policies, or add their own checkers without touching the core binary.
Editorial management
The private editorial site covers the full lifecycle of a submission:
- Structured submissions with file upload, metadata capture, and author ORCID verification.
- Reviewer recruitment — invite reviewers by email; each gets a private review workspace. An AI-powered recommender suggests candidates based on manuscript content.
- Workflow tracking — configurable status transitions (submitted → under review → accepted → published) with a full audit log.
- Messaging — compose templated messages to authors and reviewers from within the editorial interface; reviewer forms submit inline without a page reload.
- Role-based access — editors, managing editors, authors, and reviewers each see only what they need.
AI-assisted copyediting and typesetting
- LLM proofreading for typos and grammar.
- Friendly web interface for users to review copy edits
- Automatic rendering to HTML, PDF, and JATS
Public journal site
jiab publish generates a complete static website from accepted articles — no separate CMS required. The site includes full-text article pages, a browsable archive, volume and issue navigation, an Atom feed, and a client-side search index. Navigation pages (aims, editorial board, submission guidelines) are plain Markdown files; no template expertise needed.
Post-publication review
Published articles support a moderated comment thread visible on the public site. Editors review each comment before it appears. An LLM-assisted moderation step pre-screens submissions for relevance, civility, and spam, reducing the manual workload for small editorial teams. Comments become part of the permanent scholarly record attached to the article.