Tool
Find HTML and wiki scrape leftovers
HTML tags, named or numeric entities, [[Category:]]/[[File:]] wiki links, and markdown links are matched as patterns. Stripping is substitution, not an HTML parser. Nothing leaves this tab.
Pattern matches run in this tab. The paste is not uploaded.
Which leftovers are matched
HTML tags: a conservative </?name …> pattern. It is not a browser parser and will miss broken markup.
Entities: &name; , &#digits; , &#xhex; . Only nbsp, lt, gt, amp, quot, apos, and numeric forms are decoded when stripping. Unknown names are left as written.
Wiki: [[분류:]] [[파일:]] [[틀:]] and the English Category/File/Image/Template prefixes are namespace hits. Other [[link|text]] pairs keep the text after the last pipe when stripping.
Markdown: [label](url) and . The label or alt is kept. JSONL mode rewrites string fields only.
What this page does not do
- It does not run JavaScript, CSS layout, or a DOM. Style blocks and unmatched < are not a full HTML tree.
- It does not decode every HTML named entity. Leaving an unknown &foo; is intentional.
- A clean strip on one line is not a finding about a whole dump. At most 20,000 lines or about two million characters.