Markdown is the writing format behind most documentation, READMEs, and technical blogs, but it is easy to ship content that is inaccessible once it renders to HTML. Missing alt text, headings that jump levels, and raw URLs all create real barriers for screen-reader users. This checker scans your Markdown source in the browser and flags those issues line by line so you can fix them before publishing.
How it works
The tool walks your Markdown one line at a time and applies four rules:
- Empty alt text — image syntax
with nothing between the brackets is flagged, because the image has no text alternative. - Skipped heading levels — it tracks the current ATX heading depth (the number of leading
#characters) and flags any heading that is more than one level deeper than the previous one. - Bare URLs — a line containing a raw
http(s)://link that is not already wrapped in[text](url)link syntax is flagged, since it will render as unreadable link text. - Header-less tables — a pipe table whose first row is not followed by a
|---|separator row is flagged, because it lacks the header row that assistive tech relies on.
Each finding shows the line number and the offending text so you can jump straight to it.
Tips and example
Consider this snippet:
## Overview
#### Details

See https://example.com/specs for more.
The checker reports three issues: the #### heading skips H3, the image has empty alt text, and the bare URL should be [the specs](https://example.com/specs). A fixed version uses a ### heading, , and descriptive link text — clean across all four rules.