Platforms

How to save a Substack post as Markdown

Updated October 14, 2026 · 3 min read

Substack is built to be read in an inbox or in its app, and it shows when you try to keep a post. A reader has no export. Copy-paste brings the subscribe buttons along and loses the footnotes. To save a Substack post as Markdown, with its images and notes in the right place, you need a converter that knows how a Substack page is put together.

What Substack offers

  • For writers: an export of the publication from its settings, which delivers the author's own posts in a ZIP archive. It is a migration tool, and it only covers what you published.
  • For readers: a Save button, which keeps a bookmark inside Substack. There is no download.

If the newsletter you want to keep is not yours, the page is all you have.

Save a Substack post to Markdown with Markforge

Markforge is a free Chrome extension that converts the page you are on to Markdown, with a parser written for Substack posts.

  1. Install Markforge from the Chrome Web Store.
  2. Open the post in your browser, signed in if it is a paid post you subscribe to.
  3. Click the Markforge icon, or press Alt+M.
  4. Click Copy or Download .md.

The parser recognizes a Substack post on substack.com and on custom domains. It keeps the article and removes what surrounds it: subscribe boxes, share and like buttons, comments, recommendations.

What the Markdown file looks like

We ran the parser on a public post from Astral Codex Ten, a Substack publication on its own domain. The top of the file, with the Add title as H1 heading setting turned off:

---
title: "Your Book Review: We Present Television"
author: "Scott Alexander"
date: 2026-10-10
source: "www.astralcodexten.com"
url: "https://www.astralcodexten.com/p/your-book-review-we-present-television"
description: "Finalist #11 in the Book Review Contest"
---

# Your Book Review: We Present Television

*Finalist #11 in the Book Review Contest*

By **Scott Alexander** - October 10, 2026

* * *

The subtitle becomes the description in the frontmatter and an italic line under the title. Author and date are read from the page.

On that post, the article body counts about 9,100 words. The export came out complete, with its 11 images and 11 footnotes. In tokens, the page went from 114,392 as HTML source to 14,338 as Markdown.

Images and footnotes

Images are kept at their position in the text. By default Markforge downloads them into an images folder next to the Markdown file, so the note survives if the post is taken down. Turn that off in the settings to keep them as links to Substack's servers.

Footnotes keep their numbers. The markers in the text are links, and the notes are listed at the end of the file.

Video and audio embeds are removed. The text around them is kept.

Sending posts to Obsidian or to an LLM

To build a library of the newsletters you read, set a folder of your vault as the download folder. Each post becomes a note with author, date and source as properties. The setup is described in saving AI conversations to Obsidian, and it is the same for any page.

To discuss a post with a model, click Copy and paste. Long essays are long: check the token count in the popup, and use Strip links to drop the URLs if the essay cites heavily.

Limits worth knowing

  • The title appears twice with default settings. The parser writes the title, and so does the Add title as H1 heading setting. Turn that setting off, or delete one of the two lines in the preview.
  • Paid posts need your subscription. Markforge exports what your browser shows. On a locked post it exports the preview and adds a note that the content may be truncated.
  • Tables are flattened. Version 1.0.0 writes each cell on its own line.
  • Posts only. The parser matches post pages, whose address contains /p/. Notes and the publication's home page go through the generic parser.
  • Chrome only.

For other long-form pages, see how to copy a web page as Markdown for an LLM. To compare Markforge with other clippers, including those better suited to plain articles, read the best Markdown web clippers for Chrome.

Frequently asked questions

Can I export a Substack post I am reading?

Not from Substack itself. Its export tool is for publication owners and covers their own posts. As a reader, you need to convert the page, for instance with a browser extension.

Does it work on Substack publications with a custom domain?

Yes. Markforge recognizes a Substack page by its markup, not only by a substack.com address. We tested it on a post published on a custom domain on October 14, 2026.

What happens with paid posts?

Markforge converts what your browser displays. Signed in as a subscriber, you get the full post. On a preview, you get the preview, and a note at the end of the file says the content may be truncated.

Are footnotes kept?

Yes. Footnote markers stay in the text as numbered links, and the notes themselves are listed at the end of the file.

Are the comments exported?

No. The parser removes the comments section, the subscribe prompts, the share buttons and the recommendations, and keeps the article.

Related guides

How to copy a web page as Markdown for ChatGPT or ClaudePasting a page into a model works better as Markdown than as HTML or raw text. Three ways to get it, and which one works on pages behind a login.How to save ChatGPT, Claude and Gemini conversations to ObsidianPoint the export at a folder in your vault once. After that, each ChatGPT, Claude, Gemini or Grok conversation becomes a note with properties in one click.The best Markdown web clippers for Chrome in 2026Six maintained Markdown web clippers, compared on where the output goes, what happens to images, and how they handle pages that are not articles.