Almost every scholarly journal article that ends up in PubMed Central, an institutional repository, or a publisher’s own full-text database passes through JATS XML at some point in its production. It is the standard that defines how the article’s text and metadata get structured into machine-processable markup — not a specific journal’s format, but a shared vocabulary that lets publishers, archives, and downstream systems exchange the same article without each side building bespoke parsing logic. If you work in a research office, editorial team, or repository and have ever seen a deposit requirement that says “JATS-tagged XML,” this guide explains what that actually means.
What JATS is
JATS (Journal Article Tag Suite) is an XML vocabulary, standardized as ANSI/NISO Z39.96, for representing the content and metadata of a scholarly journal article — title, authors and affiliations, abstract, funding and licensing information, the full body text with its section structure, and the reference list — in a single structured document. It grew out of the National Library of Medicine’s Journal Archiving and Interchange DTD, first developed in the early 2000s for PubMed Central, and was formalized as a NISO standard (first approved 2012). The current published version is ANSI/NISO Z39.96-2021 (JATS version 1.3), maintained by NISO with the working Tag Library published at jats.nlm.nih.gov, the primary reference for anyone implementing it.
The key idea is separation of structure from presentation: a JATS document doesn’t say “put this in bold, 14pt type,” it says “this is the article title,” “this is a figure caption,” “this is a funding statement.” Systems downstream — a PDF renderer, an HTML article-view page, a text-and-data-mining pipeline, an archive’s long-term preservation store — each apply their own presentation or processing to the same underlying tagged content.
The three tag sets: Green, Blue, and Pumpkin
JATS is not one fixed schema but a suite of three related tag sets, each intended for a different point in an article’s lifecycle. They’re commonly referred to by the color of their reference documentation:
- Archiving and Interchange (nicknamed Green) — the most permissive tag set, designed for long-term archives and interchange between systems. This is the tag set PubMed Central (PMC) uses as its archival format: PMC transforms the full text it receives into XML valid against the JATS Archiving and Interchange DTD, producing what PMC refers to internally as NXML.
- Journal Publishing (nicknamed Blue) — a somewhat more constrained tag set intended for publishers depositing content with archives or aggregators, with tighter rules than Green about what a valid document must contain.
- Article Authoring (nicknamed Pumpkin) — the most restrictive of the three, designed for authoring tools and conversion pipelines working with a manuscript before it has been through full editorial and production tagging.
A publisher’s production system typically doesn’t touch a single tag set exclusively — a manuscript might be converted through an Authoring-tag-set-compatible intermediate form, tagged for Publishing at the point of typesetting, and then transformed again into Archiving-compliant XML for deposit with PMC or another archive.
What a JATS document actually contains
Every JATS article document is organized into three top-level parts:
- Front matter — the article’s metadata: title, author names and affiliations, ORCID iDs where supplied, abstract, keywords, funding information, publication dates, journal-level metadata (ISSN, publisher), and licensing/copyright statements.
- Body — the full text itself: sections and subsections, paragraphs, figures, tables, equations, and their captions, tagged with structural elements rather than visual formatting instructions.
- Back matter — the reference list, appendices, and supplementary material notes, each reference broken into its own structured citation elements (authors, title, journal, volume, pages, DOI) rather than left as a single unparsed text string.
Because references, funding statements, and author affiliations are tagged as discrete structured elements rather than free text, a JATS document is directly machine-actionable: a citation graph, an ORCID-linked author list, or a funder-compliance report can be extracted programmatically without natural-language parsing of the article’s prose.
Why JATS exists, and who actually uses it
Before a shared tagging standard, every publisher’s XML (where XML existed at all) followed its own internal schema. An archive like PMC receiving full text from hundreds of different publishers, each with a different proprietary structure, faced an unsustainable amount of custom per-publisher conversion work. JATS exists to solve exactly that problem: one shared, well-documented vocabulary that any conforming system can produce and any conforming system can consume, without publisher-specific integration work.
In practice, JATS is used by:
- PMC and Europe PMC, which require content in (or convert content into) JATS-conformant XML as their archival and full-text-delivery format.
- Journal publishers, who use it as a production/typesetting intermediate format, a deposit format for archives, or increasingly as an authoring target within submission and production systems.
- Aggregators, discovery platforms, and full-text repositories — for example Erudit, a major Canadian scholarly publishing platform, uses JATS XML as its underlying full-text infrastructure.
- Text-and-data-mining (TDM) pipelines and research tooling, which rely on JATS’s structured markup to reliably extract sections, references, and metadata at scale across articles from many different publishers — something that would be far less reliable against inconsistent per-publisher HTML or PDF.
Two sibling tag suites extend the same underlying design philosophy to adjacent content types: BITS (Book Interchange Tag Suite) for book and book-chapter content, and STS (Standards Tag Suite) for technical standards documents.
JATS XML vs. Crossref deposit XML: two different, complementary schemas
It’s easy to conflate JATS with Crossref deposit XML, since both are XML formats associated with journal articles, but they serve different purposes and aren’t interchangeable:
- JATS tags the article’s full text and structural content — the actual body, section by section — for archiving, interchange, and processing.
- Crossref deposit XML is a bibliographic registration format: publishers submit it to Crossref specifically to register a Crossref DOI and associated citation metadata (title, authors, journal, volume/issue/page, funding data, reference list) — it does not carry the article’s full body text the way JATS does.
A publisher’s production pipeline commonly produces both from the same underlying content, for different destinations: JATS XML for archival deposit (e.g. to PMC), Crossref XML for DOI registration and citation-metadata deposit. They’re complementary layers of the same publication event, not competing standards, and a well-built production system generates both rather than treating one as a substitute for the other.
How JATS integrates with machine-readable access and licensing metadata
The JATS Tag Library also formally defines the elements used by NISO Access and License Indicators (ALI), NISO RP-22 — specifically the ali:free_to_read element (a machine-readable access flag, independent of licensing) and ali:license_ref element (a URI pointing to the actual applicable license). These live in the ali: namespace within the JATS front matter, letting a system determine an article’s access status and reuse terms directly from the tagged metadata rather than having to parse a human-readable rights statement in prose. This is a concrete example of why the underlying tag suite matters beyond archiving alone: metadata standards like ALI are built to extend JATS rather than to replace or duplicate it.
Where JATS sits relative to other metadata standards
JATS is one part of a broader ecosystem of scholarly-metadata schemas, each covering a different layer. DOAJ (the Directory of Open Access Journals), for instance, considered but ultimately did not fund broader native JATS acceptance for its own inbound metadata submissions; instead, alongside its own DOAJ Native XML schema, DOAJ accepts Crossref XML as one of its supported inbound formats — a practical example of publishers reusing the Crossref deposit format they already produce, rather than adding JATS output specifically for DOAJ indexing. This illustrates a broader point: which schema a given system asks for depends on what that system actually needs (full text vs. bibliographic registration vs. index-level discovery metadata), not on any single format being universally required everywhere.
Why this matters if you’re not a publisher
Research administrators, institutional repository managers, and publishing-office staff run into JATS indirectly more often than they might expect:
- PMC deposit compliance. When a funder mandate (e.g. an NIH Public Access Policy deposit) requires a paper to appear in PMC, the underlying mechanism is that the publisher or author manuscript gets converted into JATS-conformant XML as part of that deposit — understanding this helps explain why some deposit routes (publisher-direct XML submission) are more reliable than others (manuscript conversion from a PDF).
- Evaluating a production or repository vendor. A system’s ability to import, export, or validate JATS XML is a genuine, checkable capability distinction between platforms, not vendor marketing language.
- Understanding TDM and analytics tooling. Institutional bibliometrics, evidence-synthesis, or text-mining projects that pull structured full text at scale are very often working against JATS-tagged corpora (e.g. the PMC Open Access Subset) specifically because the structure is reliable enough to process programmatically.
Common misconceptions
- “JATS is a file format.” It’s a tag vocabulary (a schema/DTD family) for XML, not a file extension or a single fixed document type — a JATS document is just XML that conforms to one of the three tag sets.
- “JATS and Crossref XML are the same thing.” They aren’t interchangeable, see above — one tags full-text content for archiving, the other registers bibliographic metadata for DOI assignment.
- “Authors need to write in JATS.” Authors write and submit manuscripts in ordinary formats (Word, LaTeX); JATS tagging happens during production/typesetting or at archival deposit, generally invisible to the author.
- “All journals produce JATS.” Adoption is wide but not universal — smaller or non-PMC-participating journals may never produce JATS-conformant XML at all, relying instead on PDF or HTML as their only full-text output.
Frequently asked questions
What does JATS stand for?
Journal Article Tag Suite — standardized as ANSI/NISO Z39.96, currently at version 1.3 (Z39.96-2021).
Is JATS XML required for PubMed Central deposit?
PMC’s archival format is JATS-conformant XML (the Archiving and Interchange, or “Green,” tag set) — content submitted to PMC is transformed into this format regardless of what format the publisher or author originally supplied, though publisher-direct XML submission is generally more reliable than conversion from a PDF or manuscript file.
What’s the difference between JATS and Crossref XML?
JATS tags an article’s full-text content and structure for archiving and interchange. Crossref deposit XML registers an article’s bibliographic metadata to obtain a Crossref DOI. They cover different layers of the same publication and are typically both produced from the same underlying content, not substitutes for each other.
Do I need to know JATS to work in research administration?
Not in the sense of writing or editing it directly — that’s a publisher/production-system task. Knowing what it is helps in evaluating repository or production vendors, understanding why PMC deposit requirements work the way they do, and interpreting text-and-data-mining or bibliometrics tooling that operates on JATS-tagged corpora.
What are the three JATS tag sets called?
Archiving and Interchange (Green), Journal Publishing (Blue), and Article Authoring (Pumpkin) — named for the color of each tag set’s reference documentation.
Related CASRAI resources
- JATS XML (dictionary term) — concise operational definition and canonical references
- NISO Access and License Indicators (ALI) — the machine-readable access/license elements defined within the JATS Tag Library
- Crossref deposit XML — the complementary bibliographic-registration schema
- Crossref DOI
- Crossref vs. DataCite
- CRediT in JATS XML — how contributor-role metadata is encoded within JATS
