Why Every Published Paper Needs XML, JATS and HTML Versions
Learning how to create XML, JATS and HTML full text for a published paper is the step that turns a journal from "a website with PDFs" into a proper scholarly publisher. A PDF reproduces the printed page well, but search engines, screen readers, archives, text-mining tools and mobile readers all work better with structured text. JATS XML gives every title, author, affiliation, heading, figure and reference a defined tag; HTML presents that same content as a readable, linkable web page.
Funders, repositories and archives increasingly expect machine-readable full text, and some services that collect articles prefer or require JATS. Even where it is optional, publishing HTML and XML improves discoverability, accessibility and long-term preservation. This guide covers the formats, the standard, practical production workflows for small and large journals, the meta tags Google Scholar reads, validation, equations and images, and accessibility.
Write the content once, tag it properly, and let every output, from PDF to HTML to archive copy, come from the same source.
PDF, HTML, XML and JATS: What Each Format Does
People often use "XML" and "JATS" as if they meant the same thing, and confuse both with the XML file sent to Crossref. They are related but have different jobs.
| Format | What it contains | Who uses it | Required? |
|---|---|---|---|
| The typeset article as fixed pages | Readers, printers, citation managers | Standard for almost every journal | |
| HTML full text | The article as a web page with headings, figures, tables and linked references | Readers on any device, search engines, screen readers | Strongly recommended |
| JATS XML | Full article content and metadata tagged to the NISO JATS standard | Archives, repositories, aggregators, text-mining tools, conversion tools | Required by some archives and services; recommended for all |
| Crossref deposit XML | Bibliographic metadata, DOI, URL, references, licence, funding | Crossref and the services that read its metadata | Required to register a Crossref DOI |
What Is JATS? The NISO Z39.96 Standard and Its Tag Sets
JATS, the Journal Article Tag Suite, is an XML vocabulary for describing journal articles. It grew out of the US National Library of Medicine's article DTDs and is now maintained by NISO as ANSI/NISO Z39.96. The current release, JATS 1.4, was approved as Z39.96-2024 in October 2024 and is kept under continuous maintenance. Documents valid under earlier versions remain valid under 1.4 of the same tag set, so moving up is low risk. The official documentation and tag libraries are hosted at jats.nlm.nih.gov, and the standard itself is listed on the NISO JATS committee page.
Archiving and Interchange
The most permissive set. Designed to capture almost any existing article for archives and exchange between systems, even when the source was not built to a strict model.
Journal Publishing
Moderately prescriptive. Intended for publishers creating new content; it guides consistent tagging and is the usual choice for a journal's own production.
Article Authoring
The most prescriptive set, for writing articles directly in XML. It allows fewer choices so that authoring tools stay simple.
Related standards worth knowing
BITS, the Book Interchange Tag Suite, extends the same ideas to books and chapters. JATS4R (JATS for Reuse) is a NISO working group that publishes recommendations on tagging specific items consistently, such as licences, data citations, authors' contributions and peer review materials. Following JATS4R advice makes your XML easier for others to reuse.
Anatomy of a JATS Article: Front, Body and Back
Every JATS article has the same top-level shape. Knowing it helps you check conversion output and talk to typesetters.
| Part | Important elements | Holds |
|---|---|---|
| <front> | <journal-meta>, <article-meta> | Journal title, ISSN, publisher; article title, contributors, affiliations, dates, volume, issue, pages, DOI, licence, abstract, keywords |
| <body> | <sec>, <title>, <p>, <fig>, <table-wrap>, <disp-formula> | Sections, paragraphs, figures with captions, tables, equations, lists |
| <back> | <ack>, <ref-list>, <app-group>, <fn-group> | Acknowledgements, references, appendices, notes, funding details |
| Identifiers | <article-id pub-id-type="doi">, <contrib-id contrib-id-type="orcid">, <institution-id> | DOI, ORCID iDs, ROR IDs for institutions |
| Licence | <permissions>, <license>, <ali:license_ref> | Copyright statement and a machine-readable licence URL such as CC BY 4.0 |
A minimal skeleton
<article article-type="research-article" dtd-version="1.4" xml:lang="en"
xmlns:xlink="http://www.w3.org/1999/xlink"
xmlns:mml="http://www.w3.org/1998/Math/MathML">
<front>
<journal-meta> ...journal title, ISSN, publisher... </journal-meta>
<article-meta>
<article-id pub-id-type="doi">10.xxxx/example</article-id>
<title-group><article-title>Paper title</article-title></title-group>
<contrib-group> ...authors, ORCID, affiliations... </contrib-group>
<pub-date> ...</pub-date> <volume/> <issue/>
<permissions> ...licence... </permissions>
<abstract> ... </abstract>
</article-meta>
</front>
<body> <sec><title>Introduction</title><p>...</p></sec> </body>
<back> <ref-list> ... </ref-list> </back>
</article>This is an outline, not a complete valid file. Real articles also declare the correct DOCTYPE or schema for the chosen tag set, and the DOI shown is a placeholder.
Tag references properly
References are where most of the value lies and where most conversion errors happen. Tag each reference with <element-citation> or <mixed-citation> and, where possible, mark authors, year, article title, source, volume, pages and DOI as separate elements. Well-structured references feed citation linking, Crossref's cited-by data and reference checks.
Crossref Deposit XML vs JATS: Two Different Files
A common misunderstanding is that registering a DOI means "uploading the XML". The file sent to Crossref follows Crossref's own metadata deposit schema. It describes the article, but it does not carry the full body text.
| Question | Crossref deposit XML | JATS XML |
|---|---|---|
| Purpose | Register or update a DOI and its metadata | Represent the complete article |
| Standard | Crossref metadata deposit schema (5.5.0 currently recommended) | ANSI/NISO Z39.96 (JATS 1.4 current) |
| Body text | No; only metadata, references, abstract and links | Yes; sections, figures, tables, equations |
| Abstracts | Accepted, tagged with JATS abstract elements | Native |
| Full-text location | Resource URL plus optional full-text links, including those used by Similarity Check and text mining | The file is itself the full text |
| Sent to | Crossref | Your website, archives, repositories and aggregators |
Crossref's schema library notes that versions earlier than 5.3.1 are scheduled for deprecation by 2028, and newer versions add support such as ROR institution IDs and CRediT contributor roles. Many publishing systems generate the Crossref deposit directly from the JATS front matter, which is another reason to keep your JATS metadata accurate. See the Crossref schema library for current versions.
Four Ways to Produce JATS XML and HTML Full Text
There is no single correct workflow. The right one depends on your volume, budget and in-house skills.
XML-first production
The accepted manuscript is converted to JATS at the start of production. Copyediting happens in the XML, and PDF and HTML are generated from it. Best quality and consistency, but needs tools and trained staff.
Word-to-JATS conversion
Converters such as Pandoc, meTypeset or docxToJats turn a styled Word file into JATS. Output must be checked by hand, especially references, tables and equations.
Outsourced typesetting
A production vendor delivers PDF, HTML and JATS as a package. Predictable, but specify the tag set, validation and accessibility requirements in the contract.
PDF-to-HTML and JATS extraction
Text and structure are extracted from the final PDF after publication. Fast for back issues, but quality depends on the PDF; always run quality checks and keep the PDF as the version of record.
Open-source tools to know
- Pandoc converts between many formats and has a JATS writer; it works well with consistently styled Word or Markdown sources.
- meTypeset converts .docx to JATS through an intermediate TEI stage.
- docxToJats is a PHP library tested with files from Word, LibreOffice and Google Docs.
- Texture is a visual editor that reads and writes JATS, useful for correcting converted files.
- XSLT stylesheets transform JATS into HTML; the NLM and community provide starting points listed on the NISO JATS wiki.
- Style the manuscriptApply template styles during copyediting
- ConvertRun the converter to JATS
- Enrich metadataAdd DOI, ORCID, ROR, dates, licence
- ValidateCheck against the DTD or schema and JATS4R
- RenderGenerate HTML and PDF from the XML
- PublishPost files and deposit metadata
Building an HTML Full-Text Article Page Readers and Search Engines Trust
The HTML page is the public face of the article. It should let a reader understand the paper without downloading anything, and let a machine identify every part of it.
What a good HTML full-text page includes
- Its own permanent URL, one article per page, linked from the issue table of contents.
- The article title, authors with ORCID links, affiliations, and dates received, accepted and published.
- A visible citation line with the journal name, volume, issue, pages and the DOI as a link.
- The full abstract and keywords near the top, never hidden behind a click.
- Real heading structure (one h1, then h2 and h3 in order) that mirrors the article's sections.
- Figures as images with captions and meaningful alt text; tables as real HTML tables, not pictures.
- References as a numbered list, each with a DOI link where one exists.
- The licence statement with a link, for example to CC BY 4.0, and links to the PDF and JATS XML.
- A "how to cite" box and correction or retraction notices if any apply.
Structured data beyond meta tags
Many journals also add schema.org ScholarlyArticle data in JSON-LD, describing the headline, authors, date, publisher, licence and DOI. General search engines can use it, and it is a useful cross-check that your metadata is complete. It does not replace the citation_ tags that Google Scholar documents, so publish both, generated from the same source so they never disagree.
Supplementary files and data
Link datasets, code and extra figures from the article page with clear labels, file types and sizes. In JATS, describe them with <supplementary-material> and, where data sit in a repository, cite them in the reference list with their DOI so they can be credited and tracked.
Performance and longevity
Keep pages light: compressed images, no heavy scripts needed to show the text, and readable on a phone. Avoid building the article text in JavaScript; some crawlers and archive tools will see an empty page. If you redesign the site, keep the old article URLs working with permanent redirects rather than sending them to the home page.
citation_ Meta Tags: Making HTML Full Text Visible in Google Scholar
Google Scholar reads bibliographic meta tags in the head of each article page. Its inclusion guidelines support Highwire Press tags (the citation_ family), bepress and PRISM tags, and treat Dublin Core as a last resort because it lacks clear fields for journal title, volume, issue and pages.
| Tag | Value | Status |
|---|---|---|
| citation_title | Article title | Required |
| citation_author | One tag per author, name only | At least one required |
| citation_publication_date | Date as YYYY/MM/DD, or the year alone | Required |
| citation_journal_title | Full journal name | Recommended |
| citation_issn | Journal ISSN | Recommended |
| citation_volume, citation_issue | Volume and issue numbers | Recommended |
| citation_firstpage, citation_lastpage | Page range or article number | Recommended |
| citation_doi | The article DOI | Recommended |
| citation_pdf_url | Absolute URL of the PDF | Recommended when a PDF exists |
| citation_author_institution | Affiliation, placed after the author it belongs to | Optional |
Example
<meta name="citation_title" content="Paper title">
<meta name="citation_author" content="Sharma, Priya">
<meta name="citation_author" content="Rao, Arjun">
<meta name="citation_publication_date" content="2026/10/15">
<meta name="citation_journal_title" content="Journal of Modern and Versatile Research Approach">
<meta name="citation_issn" content="2984-8911">
<meta name="citation_volume" content="...">
<meta name="citation_issue" content="...">
<meta name="citation_pdf_url" content="https://jmvra.org/.../paper.pdf">The author names and values in this example are placeholders for illustration.
Rules from Google Scholar's inclusion guidelines
- Each article and abstract needs its own file and unique URL; several papers on one page are not indexed well.
- The full text or the complete abstract must be visible without logging in, clicking, dismissing pop-ups or scrolling through barriers.
- The title, first author's full name and publication year are the minimum; pages missing them are treated as having no tags.
- citation_pdf_url must be an absolute URL, and Scholar asks that the file sit in the same subdirectory as the HTML abstract page.
- PDFs need searchable text and each file must not exceed 5MB.
- Every article should be reachable from the home page within a few simple HTML links, and robots.txt must not block article or browse pages.
Read the full rules at Google Scholar's inclusion guidelines for webmasters.
Validating JATS XML and HTML Before Publication
An XML file can look fine in a browser and still be invalid. Validation catches errors before archives or partners reject the file.
- Check 1: Well-formed
Every tag closed, attributes quoted, special characters escaped. Any XML editor or parser will report this.
- Check 2: Valid against JATS
Validate against the DTD, XSD or RELAX NG schema of your chosen tag set and version.
- Check 3: Best practice
Run the JATS4R validator, which applies Schematron rules for its published recommendations.
- Check 4: Partner rules
If you send content to an archive with its own tagging guidelines, use its style checker; for example, the PMC Style Checker tests conformance to PMC's guidelines.
- Check 5: HTML
Run the W3C markup validator and an accessibility checker on the rendered page.
- Check 6: Content match
Compare the XML and HTML with the PDF: title, authors, abstract, every heading, figure count, table count and reference count.
Equations, Images and Tables: MathML and Figure Handling
Equations with MathML
JATS carries mathematics in <inline-formula> and <disp-formula> elements, ideally containing MathML in the mml namespace. MathML can be rendered in modern browsers and read aloud by screen readers, and it can be searched and reused. Equations saved only as images cannot be read by assistive technology; if an image fallback is unavoidable, add alt text that spells out the expression, and also keep a TeX version inside <tex-math> if your pipeline supports it.
Figures
- Keep high-resolution originals for the PDF and web-optimised versions for HTML.
- Tag each figure with <fig>, a label, a caption and <alt-text> or <long-desc> where needed.
- Name files predictably, for example by article ID and figure number, so links never break.
- Check the licence of reused figures and state permissions in the caption.
Tables
Tag tables as structured data with <table-wrap> using the XHTML-based table model, with header cells marked. A table pasted as an image cannot be searched, resized or read by a screen reader. For very large data tables, publish the data as a supplementary file and link it.
Illustrative ranking, not measured data.
Accessibility: Full Text Everyone Can Read
Accessible full text is not only a legal or policy issue; it reaches more readers. The international reference is the W3C's Web Content Accessibility Guidelines (WCAG). Text alternatives must serve the same purpose as the image, so alt text for a chart should state the finding, not just "graph of results".
Do
- Declare the page language (lang="en") and mark passages in other languages.
- Use headings in order and descriptive link text.
- Write alt text that explains what a figure shows and why it matters.
- Use MathML for equations and real tables for data.
- Keep strong colour contrast and do not rely on colour alone in charts.
- Make PDFs tagged, with a logical reading order.
Avoid
- Scanned page images instead of text.
- Text inside images, including tables and equations, without alternatives.
- Alt text that repeats the caption or says "image".
- Layouts that break when zoomed or read on a phone.
- Links labelled "here" or "read more".
How JMVRA Publishes HTML, XML and JATS Versions of Its Papers
JMVRA, the Journal of Modern and Versatile Research Approach (online ISSN 2984-8911), publishes its papers in several forms: the PDF, an HTML full-text page and a JATS XML version built on the Journal Publishing tag set. The HTML pages carry citation_ meta tags for Google Scholar, link the PDF and XML, and show the CC BY 4.0 licence. Articles carry Crossref DOIs under the prefix 10.56975, which is shared by the group's journals.
Authors who publish with JMVRA, TIJER or JNRID therefore get a paper that can be read on screen, cited accurately and reused by indexing and archive services. Publication with JMVRA costs ₹971 for Indian authors and US$41 for international authors, payable only after acceptance.
Tools that support this workflow
- JWJMS handles online submission, the author dashboard, desk screening, reviewer invitations, revision rounds, decisions, issue building and publication scheduling, with a full record of each step, so the accepted version moves into production cleanly.
- JWDOI registers Crossref DOIs and deposits article metadata, and gives journals a sponsored DOI prefix without joining Crossref directly.
- OpenRGATE is a free open-access repository and journal-discovery platform, giving published work another place to be found.
If your journal wants help producing XML, JATS or HTML beyond these listed capabilities, contact the team to discuss your requirement at [email protected] or on WhatsApp at +91 91063 65667.
Common Mistakes When Creating XML and HTML Full Text
Calling the Crossref file "the XML"
A Crossref deposit is not full-text JATS. Archives asking for XML expect the article content.
Metadata that disagrees
Different author spellings or dates in the PDF, HTML, JATS and Crossref record confuse indexers and split citations.
One page for a whole issue
Google Scholar needs one article per page and one URL per article.
References as plain text
Untagged references lose citation linking and DOI matching.
Skipping validation
Invalid files are rejected by archives and break HTML rendering.
Fixing only one format
A correction must reach the PDF, HTML, XML and the Crossref record together.
Conclusion: Creating XML, JATS and HTML Full Text That Lasts
Creating XML, JATS and HTML full text for published papers is achievable for journals of any size. Choose the JATS Journal Publishing tag set, pick a workflow that fits your volume, whether XML-first, Word conversion, outsourcing or careful extraction from the final PDF, and make sure every output says the same thing. Add the citation_ meta tags Google Scholar reads, keep the Crossref deposit accurate, validate every file, and treat MathML, alt text and real tables as standard practice rather than extras.
The reward is a paper that readers can open anywhere, machines can understand, and archives can preserve, which is exactly what authors expect when they choose where to publish.
How Journey & Wisdom Helps
JMVRA is published within the Journey & Wisdom group, which runs four shared services — similarity checking, journal management, DOI registration and conference management — and thirteen publication platforms. These are the ones that matter most for this guide. For anything not listed here, email [email protected] or message WhatsApp +91 91063 65667 to discuss your requirement.
JWJMS
Journal management systemRuns the editorial workflow: online submission portal, author dashboard, desk screening, reviewer invitations and reminders, revision rounds, decisions, issue building and publication scheduling — with a full record of every step.
jwjms.org
JWDOI
Crossref DOI registrationRegisters Crossref DOIs and deposits article metadata, and gives journals a sponsored DOI prefix without joining Crossref directly — for journals, books, chapters and proceedings.
jwdoi.org
OpenRGATE
Repository & discoveryA free open-access repository and journal-discovery platform where finished work is deposited so it stays findable.
openrgate.orgSee every Journey & Wisdom product and platform · JW ecosystem map
How to create XML, JATS and HTML for a published paper: step by step
Choose your tag set and version
Use the JATS Journal Publishing tag set in the current version for your own articles.
Create a styled Word template
Give authors and copyeditors fixed styles for title, authors, headings, captions and references.
Pick a production workflow
Decide between XML-first, Word-to-JATS conversion, outsourced typesetting or PDF extraction.
Convert the accepted manuscript to JATS
Run your converter and correct structure, references, tables and equations.
Complete the front matter
Add DOI, ORCID iDs, affiliations with ROR IDs, dates, volume, issue, pages and the licence.
Validate the XML
Check well-formedness, validate against the JATS schema and run the JATS4R validator.
Render HTML from the XML
Use XSLT or your platform to produce an accessible article page with figures, tables and linked references.
Add citation_ meta tags
Include title, authors, publication date, journal, ISSN, volume, issue, pages, DOI and PDF URL.
Check accessibility and content match
Test alt text, headings and contrast, and compare HTML and XML with the PDF.
Publish and deposit
Post PDF, HTML and XML, then register or update the DOI metadata with Crossref.
Maintain every version
Apply corrections to all formats and the Crossref record together.
Frequently Asked Questions
What is JATS XML?
JATS, the Journal Article Tag Suite, is an XML standard (ANSI/NISO Z39.96) for describing journal articles. It tags the metadata, sections, figures, tables, equations and references of a paper so that archives, repositories, aggregators and publishing tools can read, convert and preserve it consistently. The current release is JATS 1.4, approved in 2024.
How do I create JATS XML from a Word file?
Style the manuscript consistently using a journal template, then convert it with a tool such as Pandoc, meTypeset or docxToJats, or ask a typesetter to do it. Afterwards, correct references, tables and equations by hand, add complete metadata such as DOI and ORCID, and validate the file against the JATS schema.
What is the difference between XML and JATS?
XML is a general markup language for structured data. JATS is one specific XML vocabulary, with defined element names and rules, built for journal articles. A file can be XML without being JATS; a JATS file is always XML that follows the NISO standard's tag set. When an archive or partner asks a journal for "XML", it almost always means JATS.
Is Crossref XML the same as JATS?
No. Crossref deposit XML follows Crossref's metadata schema and registers a DOI with bibliographic details, references, abstract and links. It does not carry the article's body text. JATS represents the whole article. Many systems generate the Crossref deposit from the JATS front matter.
Which JATS tag set should a journal use?
Most journals producing their own articles use the Journal Publishing tag set, which is moderately prescriptive and encourages consistent tagging. The Archiving and Interchange set is more permissive and suits archives that ingest varied content. The Article Authoring set is the strictest, for writing directly in XML.
Is HTML full text necessary if I already publish PDFs?
It is not mandatory, but it is strongly recommended. HTML reads well on phones, works with screen readers, lets readers jump to sections and references, and gives search engines more structured content. Google Scholar indexes both HTML and PDF, provided each article has its own page and the abstract or full text is visible.
Which meta tags does Google Scholar need?
At minimum, citation_title, at least one citation_author and citation_publication_date. For journal articles, also add citation_journal_title, citation_issn, citation_volume, citation_issue, citation_firstpage, citation_lastpage, citation_doi and citation_pdf_url. Google Scholar prefers these Highwire Press tags over Dublin Core. The values must match what is shown on the page and in the PDF, and a page without title, first author and year is treated as having no tags.
How do I validate a JATS XML file?
First confirm the file is well-formed, then validate it against the DTD, XSD or RELAX NG schema for your tag set and version. Next, run the JATS4R validator for best-practice checks and any partner-specific style checker. Finally, compare the content with the published PDF by eye.
How are equations handled in JATS and HTML?
Equations go into inline-formula and disp-formula elements, ideally as MathML, which browsers can render and screen readers can read aloud. A TeX version can be kept alongside. Equations stored only as images are not accessible, so if you must use them, add alt text that spells out the expression.
Can I create JATS for back issues that only exist as PDFs?
Yes. Text and structure can be extracted from the final PDFs and converted to JATS and HTML. Quality depends on how the PDFs were produced, so plan for checking references, tables and special characters, and keep the original PDF as the version of record.
Does JMVRA publish HTML and JATS versions of papers?
Yes. JMVRA publishes each paper as a PDF, an HTML full-text page and a JATS XML version, with citation_ meta tags for Google Scholar, a CC BY 4.0 licence and a Crossref DOI under the prefix 10.56975. The publication fee is ₹971 for Indian authors or US$41 for international authors, payable only after acceptance.
How do I make journal articles accessible?
Follow WCAG: declare the language, use headings in order, write alt text that conveys a figure's meaning, use real tables and MathML, keep strong contrast and descriptive links, and produce tagged PDFs. Test pages with an accessibility checker and, where possible, a screen reader.
Do I need JATS to get a DOI?
No. A Crossref DOI needs a metadata deposit in Crossref's schema, not a JATS file. However, good JATS front matter makes generating accurate Crossref deposits easier, and some archives and discovery services that collect full text expect JATS. Treat the two as partners: JATS describes the article, and the Crossref record points the world to it.
What file size limit does Google Scholar set?
Google Scholar's inclusion guidelines say each full-text file must not exceed 5MB and PDFs must contain searchable text. Compress images sensibly, avoid scanned pages and Type 3 fonts, and keep supplementary data in separate files linked from the article page. Large files may not be processed as journal articles at all.
Popular Searches on This Topic
The questions and phrases researchers and journal owners search for most on this subject — each one is answered in the sections above.
- JATS XML
- how to create JATS XML
- XML for journal articles
- HTML full text journal
- JATS NISO Z39.96
- JATS 1.4
- Journal Publishing tag set
- Journal Archiving and Interchange tag set
- Article Authoring tag set
- Crossref deposit XML
- Crossref schema 5.5.0
- Word to JATS conversion
- citation_ meta tags
- Google Scholar meta tags
- citation_pdf_url
- JATS4R validator
- PMC Style Checker
- MathML in JATS
- accessible journal articles
- alt text for figures
- XML first publishing workflow
- machine-readable full text
Publish Your Research in JMVRA
Peer-reviewed, open access (CC BY 4.0), monthly and multidisciplinary · ISSN 2984-8911 · ₹971 for Indian authors and US$41 for international authors, payable only after acceptance.