I’ve been working on Apex for a while: a unified Markdown processor that pulls together the good parts of CommonMark, GFM, MultiMarkdown, Kramdown, and Marked itself. With Marked 3.2.6, Apex lands in the app as a built-in processor (beta), so you can try that whole stack without leaving the preview.
The latest version is out now for Paddle customers. There was a bit of a hiccup with the build that went out last night, so if you’ve found yourself unable to launch Marked to check for updates, download the Free Trial from the website (your registration will already be in place). For MAS and Setapp users, the update is in review and should be available soon.
Markdown flavors keep splitting. Pick MultiMarkdown for citations and you lose something else. Pick GFM for task lists and you’re fighting table quirks from another tool. Apex’s goal is simpler: one converter that can speak those dialects, and a Unified mode that turns on the useful bits together.
In Marked, choose Apex from the processor menu, then open Configure Apex for the optional syntax toggles (tables, footnotes, wiki links, indexes, and more). There’s also a Configure Apex Files group for CSL styles, bibliography files, concordance, and metadata paths, with browse buttons and sandbox checkmarks so Marked can actually reach those files.
A lot of day-to-day Markdown pain is inconsistency: three footnote styles, two attribute list dialects, tables that “almost” parse depending on which processor wrote them. Apex leans into that mess on purpose.
It accepts multiple citation and footnote syntaxes, Kramdown- and Pandoc-style attributes, wiki links, callouts, definition lists, and the usual GFM extras. The point isn’t to invent yet another dialect. It’s to stop punishing you for having files that grew up in different tools.
I eventually intend to replace all of the built-in processors with just Apex, as it will cover the vast majority of the reasons you might pick any one processor. It will include a mode selector so you can narrow down to just one feature set (CommonMark, Kramdown, etc.) for compatibility with other tools. It will reduce a lot of my customer support by Just Working™ with whatever flavor your document is written in.
Apex brings real citation support into Marked’s built-in path. Point it at BibTeX, CSL JSON, or CSL YAML bibliographies, pick a CSL style, and use Pandoc-style [@key], MultiMarkdown [#key], or mmark-style cites. The bibliography can generate from what you actually cite, and you can keep those paths in document metadata or in the Apex Files sheet so each project stays self-contained.
If you write book-ish or academic Markdown, indices matter. Apex can build them with the TextIndex / mmark-style tooling you’ve seen in the CLI, and Marked surfaces that through the Apex prefs instead of a separate pipeline.
Tables get the same “be less brittle” treatment. Alongside normal GFM tables you get relaxed tables (Kramdown-style tables without a rigid separator row), headerless tables, grid tables, colspan/rowspan helpers, and captions. Fewer hand-edits just to make a grid render. Speaking of grids, there’s even optional support for Pandoc-style grid tables.
There’s more under the hood than the Settings sheet can show in one glance:
<!--TOC-->, ``, and friends)It’s still marked beta in the app for a reason. I’m shipping it so people can lean on it early, not because every edge case is done.
Separately — and honestly this ate a lot of the same release cycle — paginated PDF export got a major pass.
If you’ve been fighting Swiss-cheese mid-page holes, headings parked halfway down a page, lists that jump as a single block, blank trailing sheets, or footer lines being cut off: that was the target. The print engine is revamped, packing is tighter, long paragraphs and lists fill the remaining page and split at the cut, continuation table headers repeat cleanly, and a pile of same-page hole / footer-straddle bugs are gone.
Continuous PDF got love too (full export width, fewer blank columns, no surprise blank bands from heading page breaks), but the big story for print workflows is pagination that stays dense without inventing empty pages.
Paginated PDFs in Marked 3 are a different beast than Marked 2. The system print engine that Marked 2 relied on crashes on modern OSs when you try to add headers and footers. So in Marked 3, it builds a hidden web view the size of the page dimensions, loads the document, then paginates it using JavaScript, scrolling through by page height to capture a vector PDF context of each page. This is much slower, and that sucks. But it does open the door for things like footnotes per page, which is something I’m working on next.
From the changelog, the highlights:
New
Improved
n / p jump bookmarks by document positionFixed
<!--TOC--> filling when an H3 appears before the first H2_apex_* symbolsIf you want the kitchen-sink list, it’s all in the app’s release notes and the online changelog. If you try the Apex processor, tell me where it still fights your files — that’s the fastest way to get Unified mode where it needs to be.
Check out Marked 3 at markedapp.com.
Brett is a writer and developer living in Minnesota, USA. You can follow him as ttscoff on GitHub, Mastodon, and elsewhere.
Sign up for the email newsletter, and keep up with this blog by adding it to your favorite news reader, or following it on Bluesky or Mastodon at @brettterpstra.com@web.brid.gy.