https://heckyesmarkdown.com/api/2/?[query string](heckyesmarkdown.com and
fuckyeahmarkdown.com are mirrors, you can use either
endpoint.)
Parameters are passed as a query string to the API.
Parameter keys can be shortened to just the first letter (or letters required to be unique). Use 1 for true and 0 for false (also accepts 'true' and 'false').
The url parameter is required, all other parameters
are optional (see default values).
Example:
curl 'https://heckyesmarkdown.com/api/2/?u=https://brettterpstra.com&read=0`| Parameter | Default | Description |
|---|---|---|
url |
required | Target url to markdownify |
complete |
false | Output complete HTML page with head/body tags |
format |
'markdown_mmd' | See Valid formats. |
inline |
true | Use inline link format |
import_css |
false | Embed CSS from linked stylesheets |
json |
false | Output JSON { output, html, url, title } |
link |
false | App-specific url. Link Formats |
open |
false | Open link automatically. |
output |
'markdown' | Output format |
readability |
false | Use Readability to remove cruft |
showframe |
false | Output the Marky results page |
style |
none | CSS style (HTML output) |
html |
none | If provided, process HTML |
title |
none | If provided, use instead of extracting |
The output key differs from the format key.
While format determines to what markup the HTML is
converted, the output key determines how the results will
be delivered.
| Value | Result |
|---|---|
complete |
Complete HTML document |
html |
An HTML snippet |
markdown |
Raw converted results based on
format |
url |
URL encoded version of
format |
If json=1 is specified, this key will be overridden.
If the html parameter is provided with raw HTML (url
encoded), either by GET or POST, it will use that HTML rather than
running Readability on the url. If the url
paramater is used in conjunction with html, it will be used
as the base URL for any link conversions and as the source:
keyword for the document. If a title is provided, it will
be used as the title for the document. url and
title are both optional when processing HTML.
Marky uses a custom version of Arc90's Readability. It's a little
more lax and picks up things like author blocks and occasionally share
blocks, but is more likely to include all the pertinent content
on the page. Arc90 gets confused when the page markup splits the article
content into multiple divs and picks just one. Marky attempts to prevent
that. Enable Marky's Readability by including readability=1
in the url.
Tables are converted to Markdown and formatted nicely. Tables that contain content that's not valid in Markdown tables (per PHP Extra/MultiMarkdown spec) will be compressed (lists and line breaks replaced).
Code blocks are converted to backtick-fenced code. When possible, a specified language is applied after the opening fence.
Pandoc is used for the initial conversion, which is then cleaned by
Marky. Pandoc allows many output formats, so you're not limited to just
Markdown. You can, for example, output asciidoc or a
specific flavor of Markdown, such as commonmark,
markdown_mmd, or markdown_phpextra. The
default is markdown_mmd (GitHub Flavored Markdown).
Accepted output formats:
JSON output, either through the format parameter or with
json= in the URL, will output a blob containing
url (the original URL), markup (which is the
output of whatever format is specified), html (the rendered
HTML version of the output), and title (the extracted title
of the document). If a link=X parameter is provided, an
additional link field will be included with the encoded
link.
If the link parameter is given, output will be url
encoded and turned into a link that will operate (on Mac) on specific
apps.
| Value | Result |
|---|---|
url |
Raw encoded url, no protocol |
obsidian |
obsidian://create link (Osidian) |
nv/nvalt |
nv(alt)://make url (Notational Velocity/nvALT) |
nvultra |
x-nvultra://make url (nvUltra) |
marked |
x-marked://preview link (Marked 2) |
If open=1 is included in the URL, the generated link
will be opened automatically. This will have the effect of creating a
new note in the application of choice, or previewing the result (in the
case of marked). If the result is small, a redirect header
will be sent. If it's larger than 8k, it uses a Javascript redirect,
requiring a browser window.
Marked styles can be added by
name in the style parameter. Including a style
automatically forces complete output as HTML with head and
body tags. Marked styles include:
Example:
curl 'https://heckyesmarkdown.com/api/2/?u=https://brettterpstra.com&read=1&style=swissIf you specify a URL in the style parameter, the
<link re="stylesheet"> tags from the specified URL
will be added to the HTML output of the result. This can have odd
effects as most sites style tags and classes that won't exist in the
clean output from Marky, but sometimes offers decent styling.
If the import_css option is true, then the linked
stylesheets will be pulled in and their contents embedded, as well as
any <style> tags in the target URL. This can be a
slow process and often takes multiple seconds to complete.
Marky has special handling for some sites:
| Site | Functionality |
|---|---|
| StackOverflow/Exchange | Qs, comments, and As, "accepted" highlighted |
| GitHub Repo | Outputs just the README contents for the repo |
| GitHub Gist | Formats code block with title |
| GitHub File | Formats file code with title |
More to come as needs arise.
Open current page in Marky:
javascript:(function(){var nvwin = window.open("","nvwin","status=no,toolbar=no,width=400,height=250,location=no,menubar=no,resizable,scrollbars");nvwin.document.title = "Saving";nvwin.window.location = `https://heckyesmarkdown.com/api/2/?link=nvultra&open=1&read=1&u=${encodeURIComponent(document.location.href)}`;})();