texmark 0.6.1

Write scientific articles in Markdown and convert them to journal-ready LaTeX/PDF (Copernicus, Science, AMS, ...)

texmark

Write scientific articles in markdown — and submit them to any journal.

• Instant preview. Markdown renders live in most editors (VS Code, Cursor, JetBrains, vim) and on GitHub itself, so you see your draft as you type. No more wait-on-pdflatex cycles every time you change a sentence.
• Git + GitHub as a paper backend. Branches, pull requests, issues, blame, and diffs work the way they were designed to. Collaborators review changes with the same tooling they already use for code.
• One source, any journal. texmark compiles your markdown to any of the supported LaTeX templates. Start writing first, decide on a target journal later, and switch mid-way (or after rejection) by flipping a single yaml field — no rewriting needed.
• No lock-in. The intermediate .tex is right there next to the PDF. Unplug texmark whenever you want and continue the manuscript in plain LaTeX — useful for the final polish that journals usually demand.

Installation

for development, after cloning:

pip install -e .

and soon:

pip install texmark

Example

See example.md for a sample markdown file with yaml metadata in the header.

The command to convert the markdow to tex is:

texmark example.md

And to convert to PDF

texmark example.md --pdf

For another journal, it is enough to change the journal -> template' field in the yaml metadata. For testing it is also possible to pass -jfor--journal-template`:

texmark example.md --pdf -j science -o build/example-science.pdf --tex build/example-science.tex

See the example tex and pdf results in build

Journal templates

Pick the template that matches your target journal and add it to the yaml header of your markdown:

journal:
    template: ametsoc          # template name, see table below
    options: twocol            # optional, per-template — see the docs page

template	covers	example	docs
copernicus (aliases cp, esd, ...)	Copernicus / EGU: ACP, BG, CP, ESD, HESS, NHESS, TC, ...	PDF	docs
science	AAAS Science, Science Advances	PDF	docs
ametsoc (aliases jclim, jas, mwr, bams, ...)	AMS: J. Climate, JAS, MWR, BAMS, ...	PDF	docs
arxiv (alias preprint)	arXiv preprint, generic article-class	PDF	docs
elsarticle (alias elsevier)	Elsevier: QSR, EPSL, GPC, Earth-Science Reviews, Cell, Lancet, ...	PDF	docs
agujournal (aliases agu, jgr, grl, james, wrr, ...)	AGU: JGR family, GRL, Earth's Future, JAMES, ...	PDF	docs
springernature (aliases nature, naturecomms, natclimchange, natgeoscience, scirep)	Springer Nature: Nature, Nature Communications, Nature Climate Change, Nature Geoscience, Scientific Reports, ...	PDF	docs
pnas	PNAS	PDF	docs

Each template ships a default journal.options value chosen to produce a publication-style PDF (typeset 2-column journal look, no draft watermarks). For peer-review submission you usually want to switch to the publisher's submission options (e.g. preprint,12pt for Elsevier, draft for AGU, the empty default for AMS 1.5-spacing). The per-journal docs page lists the alternatives.

Partial support only. Before submitting, you will likely need to hand-edit the final LaTeX — appendix structure, special sections (Methods Online, Extended Data, Significance, …) and journal-specific cross-reference macros are not all wired up automatically. Use the example PDF to confirm the layout looks reasonable, then run texmark --tex out.tex and finish the manuscript in LaTeX directly.

Common yaml fields (all templates)

title: "Paper title"
authors:
  - firstname: Mahé
    lastname: Perrette
    affiliation: 1              # integer index into `affiliations` (most templates)
    email: [e-mail]
  - firstname: Co
    lastname: Author
    affiliation: 2
affiliations:
  - "Alfred Wegener Institute, ..."
  - "Another Institution"
date: "2026-05-26"
bibliography: references.bib
journal:
    template: <name>            # required: picks the journal template + filters
    options: <str or list>      # optional: class options (per-template default)
collect_figures_and_tables: true       # optional: see section below
figure-width: 80%                      # optional: pandoc figure default
figure-span: full                      # optional: wraps in figure* (full text width)

Section-style metadata can also be given as markdown # ... headings. Any of # Abstract, # Acknowledgments, # Data Availability, # Appendix, # Supplementary Material, # Significance, # Capsule, # Key Points, # Plain Language Summary, # Author Contributions, # Competing Interests, # Materials and Methods, # Funding, # Highlights, # Keywords will be extracted out of the body and injected into the right LaTeX command for the target journal. The exact list each template recognises is in texmark/filters/main.py.

Collect figures and tables at the end of the document

Just add

collect_figures_and_tables: true

to your markdown yaml metadata.

Advanced: latex template

The templates are written in jinja2.

Just copy from e.g. texmark/templates/science/template.tex to your own, e.g. custom_template.tex And run again with:

texmark example.md --pdf -j science -o build/example-science.pdf --tex build/example-science.tex --template custom_template.tex

The -j journal template option (here science) is still used to set custom filters (e.g. only \cite for Science, no \citet ; extract specific sections as metadata to be injected as {{section}} instead of {{body}} etc). The machinery is defined in texmark/filters.py and can in principle be extended or copied. Two approaches are possible:

• just add more filters via the --filters command or in the yaml metadata.
• extend the existing filters in a module, e.g. custom_filter.py, that extends the filters dict from the texmark.filters module (see the source code to check the details). And then pass it via --filters-module custom_filter parameter (or custom_filter in the metadata) to prompt the texmark filter to load that module and make it available via -j your-custom-name. Note that will require you to explicitly pass --template as well. Unless you overwrite an existing filter.

Ackowledgements

This project benefited greatly from AI support to extend the initial list of supported journals and further extend this package capability.