Paste any web page URL and get clean, GitHub-flavored Markdown back — headings, links, lists and tables intact, ready for your notes or repo.
Fetch any public web page and convert it to clean Markdown — copy it or publish it as a link.
Markdown appears here.Converted — share it as a page.
Publish this Markdown as a live page in one click. No account to start.
Sometimes the source you want is already on the web — a blog post, a docs page, a wiki article — and you need it as Markdown, not as a browser tab. Copy-pasting from the rendered page loses the structure: headings flatten, links drop their targets, and tables turn into tab-separated mush. This tool takes the other path. You paste the page’s URL, dochost fetches the live HTML for you, and the markup is converted to clean GitHub-flavored Markdown — the same url to markdown job people reach for when feeding a page into an LLM, an Obsidian vault, or a static-site repo. The fetch happens server-side so there is no CORS wall to fight, and the conversion keeps headings, ordered and unordered lists, inline links, blockquotes, code blocks and tables.
Paste the full page URL into the field above (a bare domain like example.com works too — we add https:// for you).
Press Convert. dochost fetches the page HTML on the server, then converts it to Markdown in your browser, so you see the result in a second or two.
Copy the Markdown, download it as a .md file, or hit Publish to host the converted page at its own shareable link.
A researcher is building a prompt library and wants three published articles as Markdown so they can paste them into a model’s context window without all the navigation chrome and ad markup. Saving each page and running a script would mean wiring up a fetcher and a converter just for a one-off. Instead they grab the first article’s URL, paste it in, and press Convert. The right pane fills with Markdown: the <h1> becomes a single top heading, the section subheads become ##, the inline citations stay as [text](url) links, and the one comparison table comes across as a pipe table. They copy it, drop it into their notes, and repeat for the other two URLs — three clean Markdown documents in under a minute, with no script to maintain and nothing scraped beyond the single page they asked for.
“Website to Markdown”, “webpage to MD” and “URL to Markdown” all ask for the same thing: give it an address, get that page back as Markdown. Pick the case that is yours.
You want a single page of a website as a .md file.
Paste that page’s URL above and press convert. Headings, lists, links, tables and code blocks come across; scripts and styles are dropped. Navigation menus and footers are part of the page, so they come along too — delete them from the output before you save it.
You want every page of a site in Markdown.
The tool converts one URL per run and does not follow links. For a few pages, run each one and paste the results together. For a whole site, use a crawler that honours robots.txt — most sites would rather be asked than bulk-fetched.
You want a page’s content in a form a chat model reads well.
Convert it and copy the Markdown straight into the chat. Markdown keeps the structure — headings, lists, tables — in far fewer tokens than the raw HTML, and without the markup the model has to wade through.
Pages that build their content with JavaScript may come back thin: the fetch reads the HTML the server sends, not what a browser draws later.
You need the result as a file, not text on screen.
Use the download button under the output to save it as a .md file, or publish it and get a link that renders the Markdown as a formatted page.
With HTML to Markdown you already have the markup and paste it in. Here you only have a link: dochost fetches the page’s HTML on the server first (so cross-origin pages work without CORS errors), then converts that HTML to Markdown. The conversion itself is the same GitHub-flavored engine.
No. It fetches the page’s server-rendered HTML and converts that. For most articles, blogs and docs that is the full content. Pages that build their body entirely with client-side JavaScript may come across thin, because the text is not in the initial HTML.
Any public http or https page that returns HTML. Pages behind a login, on a private network, or at a local address are not reachable — the fetch is blocked for anything that is not a public web URL.
Yes. HTML tables become GitHub-flavored pipe tables, and <pre>/<code> blocks become fenced code blocks. Headings, lists, links and blockquotes are preserved; scripts and styles are dropped.
One page per run. The tool fetches exactly the URL you paste and converts that page — it does not follow links to the rest of the site. For a handful of pages, run each URL in turn and stitch the results together. For hundreds, you want a crawler, and it should respect the site’s robots.txt and rate limits.
Links gratuitos duram 7 dias com uma conta gratuita. Em um plano pago, cada página que você publica fica online permanentemente.
Ver preços