From accepted Word file to publication-ready galley
GalleyCraft ingests an accepted manuscript, converts it to structured HTML, JATS4R-compliant XML, and a journal-styled PDF — while CheckRef and RefLens validate every reference inside the proof as a publication-blocking gate.
One canonical source for HTML, JATS, and PDF. No proof is approved with unresolved references.
Four steps from acceptance to publication
The editor sends an accepted manuscript. GalleyCraft carries it the rest of the way — ingestion, proofing, validation, and export back to your platform.
Ingest the DOCX
The accepted Word file arrives from iJournalPro. XSweet converts it into a clean canonical document model. Equations and complex tables that need rescue are handled in isolation.
Structure & convert
One ProseMirror document becomes the single source of truth. HTML, JATS4R-compliant XML, and a journal-styled PDF are deterministic outputs of it — they never drift apart.
Proof with the engines
Authors correct text and answer queries. CheckRef confirms DOIs. RefLens flags source-title scope and quality. Unresolved references become production queries that block submission.
Approve & publish
Gate A: all queries resolved + all references verified. Gate B: JATS4R clean + reference gate + PDF visual-regression pass. Then JATS, PDF, HTML, and DOI return to iJournalPro.
Proofing built around publication integrity
The galley stage is not only for typos — it is where references are verified, metadata is locked, and the JATS output is validated, using the same engines you already trust on CheckRef.
One canonical source → three formats
HTML, JATS4R XML, and a print-ready PDF are all produced from a single ProseMirror document. Changing the source changes all three consistently — there is no "primary" export file to drift from.
Reference trust gate (CheckRef)
Every reference receives a CONFIRMED, UNCERTAIN, or NOT_FOUND verdict. CONFIRMED auto-satisfies the verification gate. Uncertain and not-found references require author evidence links before the proof can be submitted.
Source-title intelligence (RefLens)
RefLens checks JCR and SJR recognition, journal-title normalization, scope-fit signals, and quality flags — surfaced inline per reference in the proofing panel.
JATS4R validation as a CI gate
Every export is validated against the JATS DTD and JATS4R Schematron. Data citations receive the correct publication-type="data" tagging. The approval gate only opens when JATS4R is clean.
Production queries & metadata locking
Structural metadata (DOI, author names, affiliations, publication type) is locked from author editing. Changes are routed to production as queries. Gate A blocks submission until every query is resolved.
Journal-styled PDF visual regression
Each journal template generates a PDF through Paged.js. Automated visual-regression checks catch page-count drift, figure placement issues, table overflow, and citation-link integrity before approval.
Built for everyone in the post-acceptance workflow
From the production desk to the corresponding author, everyone works on one proof — each with the right access level.
Production editors
Hand off an accepted DOCX and receive clean JATS, HTML, and PDF with reference issues caught before the author opens the proof. Queries and metadata locks keep the production workflow in your hands.
Authors
Access the proof via a magic link — no CheckRef account needed. Review the typeset article in your journal's style, fix permitted text, answer queries, and verify references directly in the browser.
Publishers & journals
Enforce reference integrity and JATS4R quality as a publication-blocking gate, consistently across every issue you produce. Per-journal templates, policies, and validation thresholds are configurable.
One scholarly quality platform, separate product surfaces
GalleyCraft is a CheckRef platform service. Login, users, and admin are centralized at checkref.org; the proofing UI, document model, workers, and database are GalleyCraft's own.
CheckRef
Reference validation and DOI verification before and during publishing.
RefLens
Bibliography relevance and source-title intelligence for editorial teams.
GalleyCraft
Production proofing with reference integrity as a publication gate.
ScholaRef
Author visibility and journal outreach after publication.
DOI Tools
mEDRA ONIX validation, DOI activation checks, and issue exports.
What editors ask first
What is the reference verification gate?
Every reference in the proof must reach a terminal verification status before Gate A (in_proofing → submitted) opens. A CheckRef-CONFIRMED reference auto-satisfies the gate — the author only touches UNCERTAIN and NOT_FOUND references. For those, they must supply an evidence link: a DOI, publisher page, PubMed link, or (as a last resort) a manual explanation. No proof can be submitted with any reference still in the "unverified" state.
Does GalleyCraft replace our typesetting team?
No. GalleyCraft handles structural conversion (DOCX → canonical model → HTML / JATS / PDF) and validation, but it is not a design studio. Complex layout problems — equations that XSweet cannot parse, elaborate multi-page tables — are flagged for a production editor to fix before the author opens the proof. The production team remains in control of the workflow.
How does author access work?
Authors receive a proof-scoped magic link — no CheckRef account required. The link grants access only to that specific proof. Staff (production editors, editors-in-chief) use the normal CheckRef login at checkref.org/login; access is managed via the allowed_apps.galleyproof flag in the CheckRef admin panel.
What is JATS4R and why does it matter?
JATS4R (Journal Article Tag Suite for Reuse) is a community best-practice layer on top of the JATS XML standard. It ensures citations, funding statements, data-availability sections, and other elements are tagged in a way that downstream systems (PubMed, CrossRef, repositories) can reliably parse. GalleyCraft validates every export against both the JATS DTD and the JATS4R Schematron; Gate B only opens when validation passes.
Which DOCX features are supported?
The primary ingestion path (XSweet) handles standard Word body text, headings, tables, figures, footnotes, and inline math. MathType equations and complex merged tables that XSweet cannot parse are handled by an isolated rescue parser (Pandoc on the failing fragment only). Structures that still cannot be mapped are flagged for production-editor review — they are never silently dropped.
Can we use our own journal template?
Yes. Each journal has its own CSS + Nunjucks template that controls the Paged.js PDF output. Templates are versioned, so every export is reproducible from the canonical document plus the template version in use at approval time.
How does GalleyCraft connect to iJournalPro?
The production editor sends an accepted manuscript via a handoff API call (POST /api/v1/proofs) with the DOCX URL and structured metadata (authors, affiliations, funding, license). When approved, GalleyCraft posts back the JATS, PDF, HTML URLs, the article DOI, a corrections report, and a reference verification report to the iJournalPro webhook.
Ready to take a manuscript to press?
Move from an accepted Word file to a validated, publication-ready galley — references and all.