Typesetting reports to beautiful (PDF) docs: creating a generic workflow

[rufuspollock]

We need to build a reliable, elegant, and low-maintenance pipeline to convert Google Docs into publication-ready PDFs. We currently propoe using a toolchain via LaTeX - but we're open to suggestions. Other major option is manually typesetting each report using e.g. InDesign or Affinity.

Life Itself produces reports authored (usually) in Google Docs (though also could be markdown) that need to be professionally typeset and exported as high-quality PDFs, suitable for digital download -- and maybe print (e.g. via Amazon KDP or Lulu). Manual typesetting via tools like InDesign is expensive and inefficient, especially for post-publication updates. We want a scalable, reproducible pipeline—likely using Google Docs → Markdown → LaTeX → PDF—that balances automation with elegance and is usable even by non-technical collaborators. The system should output well-formatted reports and, eventually, support compiling multiple documents into a print-ready book.

See the full SCQH below

Aside: we also need to design the reports - is this here or a separate issue? 💬2025-10-09 collecting examples of elegant reports here https://www.are.na/rufus-pollock/report-inspirations-for-life-itself

Acceptance

• [ ] A working end-to-end pipeline from Google Docs to publication-ready PDF.
• [ ] Output PDF meets professional standards (typography, layout, spacing).
• [ ] Supports key formatting elements: tables, images, footnotes, references, and structured headings.
• [ ] Easily updateable when source content changes—no manual reformatting required.
• [ ] Includes clear user documentation for non-technical collaborators.
• [ ] (Optional/stretch) Ability to compile multiple reports into a single, print-ready book.

Tasks

TBC

• [ ] Evaluate Conversion Tools

  • [ ] Test tools for converting Google Docs to Markdown (e.g. gdoc2md, Docs → HTML → Pandoc).
  • [ ] Assess Pandoc’s ability to convert Markdown to LaTeX and PDF with clean output.

• [ ] Design LaTeX or Pandoc Template

  • [ ] Create/adapt a LaTeX template with clear typographic standards (fonts, spacing, headers).
  • [ ] Ensure support for footnotes, images, tables, and section numbering.
  • [ ] Match visual design to Life Itself or Second Renaissance branding (where appropriate).

• [ ] Prototype End-to-End Workflow

  • [ ] Build a functional pipeline from Google Docs → Markdown → LaTeX → PDF.
  • [ ] Automate the steps with a CLI script, Makefile, or GitHub Action.

• [ ] Document the Workflow

  • [ ] Write clear, step-by-step instructions for non-technical users.
  • [ ] Include installation and dependency setup notes.

• [ ] Test and Iterate

  • [ ] Trial the system with at least one real report (e.g. DDS white paper).
  • [ ] Collect feedback and refine layout and usability.

Shaping

SCQH

Situation

You have a growing number of reports (e.g. Second Renaissance reports, Deliberately Developmental Spaces report) authored primarily in Google Docs. You want to publish these as elegant, well-typeset PDFs—suitable for both digital download and print-on-demand distribution (e.g. via Amazon or Lulu). There’s also interest in possibly compiling these into a book in the future.

Currently, you can export directly from Google Docs to PDF, but the result is visually and structurally suboptimal. Professional typesetting (e.g. using InDesign) is available, but involves significant cost ($200–300+) and coordination overhead.

Complication

• Manual typesetting requires ongoing coordination for each report and is not scalable—any small updates (e.g. errata) require going back to the designer.
• An automated typesetting pipeline (e.g. via Markdown → LaTeX → PDF) promises reproducibility, version control, and scalability, but:
  • Conversion from Google Docs to Markdown is lossy and messy, especially with bibliographies, figures, and rich formatting.
  • Collaborators like Sylvie are less technical, which may make Markdown/LaTeX workflows burdensome unless they are abstracted or well-supported.
  • Existing automated tools (e.g. Pandoc) have limitations for report-grade design (especially layout finesse and aesthetic control).

Question

How can we develop an efficient, scalable workflow to produce professionally typeset, updateable reports—starting from Google Docs but ideally transitioning to a Markdown-based system—without incurring high coordination or technical burdens?

Hypothesis

A semi-automated pipeline based on:

• Collaborative drafting in Google Docs,
• Conversion to clean Markdown (via e.g. gdoc2md or a Pandoc filter),
• Typesetting via LaTeX with custom templates (or via Pandoc with PDF backend), can offer a low-cost, update-friendly, and scalable solution.

This setup:

• Allows non-technical collaborators to draft and edit as usual,
• Provides a single-source publishing setup to generate PDFs (and later EPUB or print books),
• Enables post-publication edits and batch formatting consistency,
• And over time, could integrate with Git-based workflows (e.g. for versioning, CI publishing, public web versions via FlowerShow or similar).

Raw

So I just want to outline quickly my need for producing nice reports, probably PDFs, starting actually from Google Docs, but probably with markdowns and intermediate preferably. And we have multiple reports at Life Itself that we've written that we would like to make beautiful, elegant. I'd also like to be able to turn them into a book, actually, that you could put out on Amazon, potentially at some point, because then people can download them that way for free or very cheaply. It's a source of potential revenue and so on. But the basic point is I've got two options. I could hire someone to typeset it in basic InDesign or Affinity, which in my experience is OK, but generally costs about minimum, I'd say, $200 to $300 a time, really, and requires a whole bunch of management. That's the issue in my experience. There's a lot of back and forth. And maybe it will get more efficient once we have a standard style and so on, and you could just dump it in. The other issue is that when you do updates, you have to go back to the person if there are any changes or errata. So the other alternative is an automated process using, I think, LaTeX to produce the PDFs. I use that to actually create the Open Revolution book relatively successfully. I haven't used it, though, to create elegant reports, and I don't know how that would work and what I would need to do to set it up for, let's say, Lulu or things like that. Yeah, but I think that there's a lot of attraction to that kind of route for me technically, though I don't know what it is like for others like Sylvie, who's less technical. So I first want to set out the situation here, which is I have reports written in basically Google Docs, now like half a dozen, the Second Renaissance reports, the Developmental Space report. And I want to produce them, create reports or PDFs for them for downloading when the report is released. And the crudest way I can do that at the moment is I just export it from Google Docs as a PDF. And that's okay, but not very professional. We could set up the headers a bit more nicely and things like that, which I haven't done, which would then output better. There are things where I could probably tweak Google Docs quite well. Yeah, I think... So yeah, the complication, I guess, is that in terms of editing... So the complication is that then producing reports is expensive. It's difficult to go back to someone when you've made a change, so like making multiple changes is kind of painful. After you've published a report, there's quite a lot of overhead in managing someone to produce the report. And the other complication, though, is if you're in Google Docs, kind of getting it into Markdown and then panned out for publications, a bit painful. If one starts out in Markdown, one's user's profit, kind of cross-referencing, it's not too bad converting across to LaTeX or something like that. But I'm concerned there's a lossiness there or painfulness, for example, in the bibliography. Essentially, that's the kind of question at the moment. Yeah, so the question for me is how to do this report publication. What's most important probably is not even the cost, but the efficiency in terms of my time and others. And I guess it is a question of scalability, like one could then produce many more reports in this kind of way. And so my hypothesis at the moment is to use a kind of Google Docs, Markdown, to LaTeX sequence, because that way I can easily re-update. I can apply this to lots of documents. I can also produce books if necessary. My question is still, yeah, and there's also an immediate need to do this. So can you kind of outline that as a situation-complication-question-hypothesis structure?

[wxs]

@rufuspollock if you wanted to skip the Google Docs phase I know some folks who use this collaborative markdown editor: https://hedgedoc.org/

[asimong]

Using Hedgedoc is a great idea, which I have suggested in the past, and would support again.

[rufuspollock]

@rufuspollock if you wanted to skip the Google Docs phase I know some folks who use this collaborative markdown editor: hedgedoc.org

@wxs @asimong

I've used hackmd (aka hedgedoc as community edition) pretty much from the start of its existence - and even tried to make it the default in Life Itself and other orgs 🤯

Whilst i love markdown, it unfortunately comes nowhere gdocs in terms of functionality especially for non-tech users e.g. in comments, suggesting (key in collaborative editing for non-tech), performance on large docs etc etc. (I also had bad experiences personally in recent times with hackmd in real time sync leading to data loss which led me to stop using it completely).

I do dream of collaborative markdown editor - though my best bet is probably obsidian with real time collaboration which is getting increasingly easy to implement. The dream of Flowershow (+ other tools) was to create an open "own your content" Notion alternative based markdown 😜 (it's getting increasingly close ...)

---

All that said. Once i have markdown i still need to render to something well typeset. Any ideas on that are gratefully received. I did my book https://openrevolution.net from markdown to latex to pdf using pandoc ... so i'm imagining something like that.