Markdown Mirror

AI tools, chatbots and crawlers increasingly prefer plain Markdown over full HTML pages — it's smaller, cleaner, and easier for them to read accurately. Markdown Mirror gives every page you choose a second, Markdown-only version at the same address with `.md` added to the end (for example, `/about` becomes `/about.md`), alongside a standard `/llms.txt` index that lists your site's pages for AI tools to discover. You don't need to understand your site's fields, content types or theme to use it — you simply tick which content should get a `.md` version, and the module reads it directly from the database and converts it for you.

Features

- A `.md` version of any page, generated on the fly from the page's own stored content — not a scrape of the rendered HTML, so no navigation, blocks, cookie banners or JavaScript leak into it.
- Opt-in by content type, vocabulary, or page. Choose whole content types and taxonomy vocabularies, enable it site-wide for Drupal Canvas pages, or opt in/exclude individual pages one path per line.
- Works on taxonomy term pages too, including a list of the content tagged with that term.
- `/llms.txt` and `/llms-full.txt`, generated automatically from whichever pages you've enabled, grouped by content type/vocabulary, each linked to its `.md` version — the emerging standard AI tools look for first.
- Configurable frontmatter (title, URL, language, content type, dates, description, tags) at the top of each `.md` page.
- No per-site setup of field names. It was built to be generalizable: it never assumes a specific field, content type machine name, or View exists, so the same install works unmodified across different sites with completely different content models.
- Safe for anonymous, public use — dangerous links and markup are stripped, unpublished/access-restricted content is never exposed through the `.md` version, and the HTML version always gets an SEO-safe `noindex` + canonical link back to itself so the `.md` copy never competes with it in search results.
- Use it when you're building an AI chatbot against your site's content, want your pages to be accurately summarized by LLM-based tools, or just want a plain-text version of your content available for any purpose.

Post-Installation

Nothing is served automatically — this is deliberate, since you should choose what AI tools can read. After installing:

1. Grant the **Administer Markdown Mirror settings** permission to the appropriate role.
2. Go to **Configuration > Content authoring > Markdown Mirror** (`/admin/config/content/markdown-mirror`).
3. Tick the content types, vocabularies and/or Canvas pages you want mirrored, or list specific/excluded pages by path.
4. Choose what appears in the frontmatter and whether `/llms.txt`/`/llms-full.txt` are served.
5. Save, then visit any enabled page with `.md` appended to its URL to see the result.

There's no new content type, text format or field to manage — it's entirely configured from that one settings page.

Additional Requirements

- Drupal 10.3+ or 11, and Drupal core's **Node** module.
- [`league/html-to-markdown`](https://github.com/thephpleague/html-to-markdown) (installed automatically via Composer).

Optional, used automatically if present:

- Core **Taxonomy**, for taxonomy term support.
- [Canvas](https://www.drupal.org/project/canvas), for Canvas page support.
- [Metatag](https://www.drupal.org/project/metatag), to use its description tag in the frontmatter.
- [Paragraphs](https://www.drupal.org/project/paragraphs), so paragraph content is followed and included.

Recommended modules/libraries

- [Metatag](https://www.drupal.org/project/metatag) — richer `description` frontmatter.
- [Canvas](https://www.drupal.org/project/canvas) — if you want Canvas pages covered as well as nodes.

Similar projects

A few other projects also convert HTML to Markdown or publish `llms.txt`, but typically require per-content-type configuration of which fields to include. Markdown Mirror is built to need none of that: it detects and extracts content generically, so the same configuration works whether your site uses Paragraphs, plain body fields, or a Canvas page with no fixed field structure at all.

Community Documentation

See the module's [README](https://git.drupalcode.org/project/markdown_mirror/-/blob/1.0.x/README.md) for a full walkthrough of configuration, Drush commands, caching behavior and troubleshooting.