Clean pages, clean Markdown.

A free tool that turns messy web pages, pasted HTML, and saved .html files into clean Markdown — stripping ads, newsletter chrome, tracking pixels, and navigation clutter in the process. It is built for researchers saving sources, writers collecting reference material, and anyone feeding documents into an LLM, a static-site blog, or a note system that works best with plain text.

Every conversion happens in your browser tab. There is no server endpoint that receives your pages, no account, and no daily run cap — the only limits are batch size and per-file size.

How it works

The pipeline runs entirely client-side in two stages. Mozilla’s Readability algorithm — the same engine behind Firefox Reader View — identifies the main article content and discards navigation, sidebars, and ad slots. Turndown then converts the clean HTML into semantic Markdown. You can paste raw page source, paste article text directly, or drag in .html, .txt, and .md files; plain text is lightly normalized and otherwise preserved verbatim.

Because browsers block cross-origin network requests, there is no URL input to fetch a page for you: open the article in your own logged-in session, copy the page source (or select all and copy), and paste it. This is also how paywalled content works — whatever you can already read in your browser can be cleaned, without sharing your session.

Output options and front matter

Five toggles shape the result. “Title as H1” prepends the extracted headline; “Keep links” preserves hyperlinks as Markdown links; “GFM tables” converts HTML tables into GitHub-flavored pipe tables; “Drop images” removes image markup for text-only vaults; and “Front matter” prepends a YAML block with the title, site name, author, and publication date — the format Obsidian, Jekyll, and Notion understand natively.

The output panel reports word count, character count, an estimated reading time, and whether the source was detected as an article, raw HTML, or plain text. You can copy the Markdown in one click, download it as a .md file, or export the same content as .txt with the Markdown syntax stripped out.

Batch cleaning and ZIP export

When you are archiving a reading list rather than one page, the batch queue takes multiple .html, .txt, and .md files and cleans them one after another, with per-file status, skipped oversize files flagged inline, and a single Download ZIP button that packages every .md result. Failed items stay in the queue so a re-run only retries what did not finish, and any single item can be opened in the editor for a closer look.

Free and Pro limits

The free tier cleans batches of 5 files with up to 10 MB per file, as many times per day as you like. Pro raises the batch to 25 files and the per-file ceiling to 50 MB. Neither tier has a run count, a monthly quota, or a watermark, and both include commercial use — a file beyond your size or batch limit is skipped with a clear notice, never truncated silently.

Quality tips

Readability is deliberately conservative: it favors the main article and may drop comments, footers, or link lists, which is usually exactly what you want. If an output misses something you need, turn the toggles off and paste the complete page source for a raw-HTML conversion. Well-formed articles from standard publishing platforms clean up best; pages built almost entirely from scripts or custom layouts may need a manual pass afterward.

Privacy

No document text passes through our servers — Readability and Turndown both run in your tab, and nothing from your pages is logged, stored, or used for training. Sources behind a login are never exposed because the page never leaves the browser that is already authenticated.