# Skill: Google Doc to Substack (including images, tables, footnotes)

Published: 2026-08-13T00:00:00.000Z
Canonical: https://wow.pjh.is/journal/skill-google-doc-to-substack

Someone asked, so here it is: [Google Doc to Substack draft](https://github.com/HartreeWorks/skill--gdoc-to-substack).

To use it, just prompt:

```
/gdoc-to-substack https://docs.google.com/document/d/...
```

What it does:

1. Fetches a Google Doc with the [gdoc CLI](https://wow.pjh.is/journal/use-gdoc-for-google-docs).
2. Opens Substack in Chrome and creates a draft post.
3. Handles details such as images, footnotes and tables.[^1]

Tested in Claude Code and Codex.

![A Google Doc containing headings, a numbered list, links, and three footnotes](https://wow.pjh.is/static/images/blog/gdoc-to-substack-source-doc.jpg)

![The same document as a Substack draft, with superscript footnote anchors and a table inserted as an image](https://wow.pjh.is/static/images/blog/gdoc-to-substack-draft.jpg)

This skill took about 15 minutes to make. Fable one-shotted the initial brief. The improvement loop would have been faster if I had prompted better—most of the time went into us exploring and rejecting an approach to rendering tables. When Fable told me that Substack doesn't support HTML tables, my first idea was "convert to ASCII", so *I directly asked to implement that*. Instead, I should have said, "Hmm, what's the best workaround here? Maybe a code block with an ASCII table?" When I recognised ASCII as a dead end, I then had the idea to do screenshots. I guess Fable would have done too, if I'd asked it to consider the options first.

[^1]: Substack doesn't support HTML tables, so the skill takes a screenshot and inserts the table as an image.