The short version
A webpage-to-Markdown converter fetches a page, removes scripts, styles, navigation and other non-content markup, and maps the remaining HTML elements onto Markdown equivalents: headings, lists, links, images, tables, blockquotes and code blocks.
Why convert a page to Markdown at all
Three reasons come up repeatedly, and they have almost nothing in common.
Documentation migration. Moving content from a CMS into a static site generator, a wiki, or a repository means getting it out of HTML and into something a human can edit. Markdown is the usual destination.
Feeding an AI system. Language models work on text, and HTML is mostly not text: a typical page is more markup than content. Converting to Markdown strips the noise while keeping the structure that tells a model what is a heading and what is a list item.
Archiving something readable. A saved HTML file depends on stylesheets and scripts that will eventually break. Markdown is plain text and will still be legible in twenty years.
What survives the conversion and what does not
Structure survives. Headings keep their level, lists keep their nesting and ordering, tables become pipe tables, code blocks keep their fencing, links keep their targets, and images keep their alt text.
Presentation does not, and that is the point. Fonts, colours, spacing, columns and anything achieved with CSS are gone. If the meaning of a page depends on its layout, Markdown will lose it.
Interactive content does not survive either. Forms, embedded players, tabbed panels and anything assembled by JavaScript after load are either dropped or reduced to whatever plain text they contained.
- Kept: headings, lists, tables, links, images, code, blockquotes
- Dropped: CSS styling, columns, fonts, colours
- Dropped: forms, embeds, tabs, anything JavaScript builds after load
- Reduced: complex nested layouts become a linear document
Pages this will struggle with
The main one is a page that assembles itself in the browser. We fetch the HTML as delivered, which for a client-rendered application is close to empty. If you get almost nothing back, that is usually why rather than a fault in the conversion.
Sites that block automated requests are the other case. Some return a challenge page or a 403 to anything that is not a browser, and we identify ourselves honestly rather than pretending otherwise.
Content behind a login cannot be reached at all, which is the correct behaviour rather than a limitation to work around.
One page at a time, and why
This converts the address you give it, not the site around it. Crawling an arbitrary site on behalf of an anonymous visitor is not something anyone can responsibly offer for free, and the abuse potential is obvious.
If you need a whole site converted, that is a different job with different tooling, and it should run with your own credentials against your own content.
For the specific case of making a site's content available to an AI agent, that is what Venbit does as a product: it crawls your site, plus your PDFs, DOCX, XLSX and CSV files, and keeps the result current.
Questions about this tool
Does it keep tables and code blocks?+
Yes. Tables become pipe tables, code blocks keep their fencing, and headings, lists, links and image alt text are all preserved.
I got almost nothing back. Why?+
Most likely the page builds its content in the browser rather than delivering it in the HTML, so there is very little for a fetch to read. Sites that block automated requests are the other common cause.
Can it convert a whole website?+
No, one page per run. Crawling an arbitrary site for an anonymous visitor is not something that can be given away responsibly.
Why is the layout gone?+
Markdown describes structure, not presentation. Headings and lists survive; columns, fonts and CSS do not. If a page's meaning depends on its visual layout, that will be lost.