MarkDownload Alternative: A Migration Guide

If you opened your browser to clip a page and found MarkDownload gone, you are not alone: as of mid-2026 its Chrome Web Store listing is no longer available, and the forum threads asking for a replacement are filling up. The good news is that choosing a MarkDownload alternative is mostly a question of what you do with the Markdown next — save an article to read later, or feed a page to an LLM. This guide maps the displaced workflow onto the tools that still ship, honestly, without pretending one option wins every job.

I will cover what is known about MarkDownload's status, which alternatives fit which job, how their features line up against what MarkDownload actually did, and how much cleaner — in tokens — the output is. I will not speculate about why the listing came down, and I will point you at the open-source project so you can keep running it from source if it is still the right fit.

What happened to MarkDownload?

MarkDownload — the free, open-source Markdown web clipper by deathau — is no longer available on the Chrome Web Store as of mid-2026, and the project has seen little maintenance for some time. The code has not vanished: MarkDownload is open-source on GitHub, so the repository and its Firefox build remain, and you can still load the Chrome version unpacked from source (covered below). What you have lost, if you installed from the store, is automatic updates and the one-click install path.

It is worth being accurate about what MarkDownload was, because it was good at its job. It ran Mozilla Readability to strip a page down to its main article, then Turndown to serialize that into Markdown — the same extraction-and-serialization lineage many converters share, including BulkMD. It saved a single clean article to a file or the clipboard, and it had a loyal following, especially among Obsidian users building a vault from web sources. None of that competence is in question; the listing is simply not there to install anymore.

What that means in practice depends on how you used it. If you clipped single articles into Obsidian, the gap is now filled by Obsidian's own clipper. If you used MarkDownload to feed pages into ChatGPT, Claude, or a RAG pipeline, a converter built for that destination is a better landing spot than another general clipper. And if MarkDownload did exactly what you needed, you can keep it running from source.

Which MarkDownload alternative fits your workflow?

The honest split is by job, not by which tool has the longer feature list. If you save single articles into an Obsidian vault, the official Obsidian Web Clipper is the natural successor. If your pages are headed for an LLM, a converter built for that job is the better fit. There is no single "best" replacement, because MarkDownload sat at the intersection of two workflows that have since specialized.

If you clip single articles into a vault

Use the official Obsidian Web Clipper. It is maintained by the Obsidian team, clips a page or a highlighted selection straight into your vault with configurable templates and properties, and is the closest like-for-like replacement for the notes use case MarkDownload served. If Obsidian is your home base and your goal is to read or reference a saved article later, this is the move — it is purpose-built for exactly that, and it removes the manual download-then-import step MarkDownload required. For a deeper walkthrough of structuring those clips, see building an Obsidian knowledge base from the web.

If you feed pages to an LLM or build a corpus

Use a converter tuned for LLM context. BulkMD keeps MarkDownload's single-page strength — one click copies the current tab as clean Markdown — and adds the parts a notes clipper has no reason to include. It preserves GitHub Flavored Markdown tables and language-tagged code fences by default, shows a live token and cost counter so you can see the payload before you paste, and sends the result straight to ChatGPT, Claude, or Perplexity. It also exports to Obsidian with YAML frontmatter and copies into Notion, so the notes path is not lost when you switch. The dividing line is bulk and tokens, not single-page quality.

How do the alternatives map to what MarkDownload did?

The clearest way to choose is to line each alternative up against MarkDownload's actual capabilities rather than its reputation. The table below maps what MarkDownload did onto the two living successors, plus MarkDownload run from source.

Read down the columns, not across the rows: the Obsidian Web Clipper wins the vault column, BulkMD wins the bulk-and-LLM columns, and MarkDownload-from-source remains a fine single-page tool if you are willing to maintain it. If you want the full feature-by-feature breakdown of converters versus reader and clipper extensions, I wrote a longer comparison on web-to-Markdown versus reader and clipper extensions.

How clean is the output, and how many tokens do you save?

The reason any of this matters for an LLM workflow is tokens, and the saving is large enough to change what you keep. Clean Markdown typically uses 60–80% fewer tokens than the same page's raw HTML, because HTML is dominated by <div> wrappers, class attributes, inline styles, and scripts that carry no meaning for the model. On a boilerplate-heavy page the reduction can reach roughly 93% as a labeled best case — one documented sample dropped from about 45,000 tokens of HTML to about 3,200 tokens of Markdown — but 60–80% is the number to plan around.

Because the token count is measured browser-side, it is tokenizer-aware: GPT-3.5 and GPT-4 use cl100k_base, GPT-4o and the o-series use o200k_base, and Claude has its own tokenizer, but the relative win holds across all of them. The full browser-side method is in estimating LLM token cost in the browser. The practical upshot for a former MarkDownload user: you are not choosing between "clean output" and "messy output" when you migrate — every option here produces clean Markdown — you are choosing whether the tool also counts tokens, converts in bulk, and hands off to a model.

As of mid-2026, the practical MarkDownload alternatives split cleanly by job: the Obsidian Web Clipper for saving single articles into a vault, and a converter like BulkMD for turning pages — one or hundreds — into low-token Markdown an LLM can read.

Can you keep using MarkDownload from source?

Yes. MarkDownload is open-source, so if it did precisely what you needed you can load it unpacked in Chrome and keep going — you simply take over its maintenance. The trade-off is honest: no automatic store updates, and you are trusting a build you assemble yourself.

1. Clone or download the repository from github.com/deathau/markdownload.
2. Install dependencies and build the extension per the repo's README (npm install, then the documented build step).
3. Open chrome://extensions, enable Developer mode in the top-right corner.
4. Click Load unpacked and select the built extension directory (the folder containing its manifest.json).
5. Pin it, and re-pull and rebuild manually whenever you want updates.

This is the right call if your workflow is single-page clipping, you already trust the source, and you do not need bulk or LLM tooling. It is the wrong call if you want something maintained that you do not have to babysit — in that case, pick one of the two living successors above.

How to migrate in about five minutes

Migrating off MarkDownload is quick because the hard part — clean extraction — is solved the same way everywhere. Decide by destination: if your clips live in an Obsidian vault, install the official Obsidian Web Clipper and point it at your vault. If your pages feed an LLM or a RAG corpus, install BulkMD from the Chrome Web Store, copy your next page with one click, and watch the token counter to confirm the payload shrank. If you genuinely only need MarkDownload, run it from source. Whichever you choose, your existing Markdown files keep working — Markdown is portable by design, which is the whole reason it was worth clipping to in the first place.