Why copied text comes out messy
Text rarely travels cleanly between applications. PDFs break every line where the page did, so a paragraph arrives as twenty short lines. Chat assistants return Markdown, so headings carry # signs and emphasis carries **asterisks**. Word processors and websites replace straight quotes with curly ones and insert non-breaking spaces that look identical to normal spaces but behave differently in search, spreadsheets and code. The cleaner fixes each of these with a separate switch, so you remove only what you want.
What each option does
| Option | Effect | Default |
|---|---|---|
| Trim spaces at line start/end | Removes indentation and trailing whitespace on every line | On |
| Collapse multiple spaces | Turns runs of spaces or tabs into one space | On |
| Remove empty lines | Deletes blank lines; paragraphs will run together | Off |
| Join broken lines (PDF) | Joins single line breaks into spaces and repairs words split with a hyphen, but keeps blank-line paragraph breaks | Off |
| Remove markdown symbols | Strips # headings, **bold**, *italic*, backticks, > quotes, list bullets and numbering, link syntax and table pipes, keeping the visible text |
On |
| Smart quotes/dashes to ASCII | “ ” → “, ‘ ’ → ’, en dash → -, em dash → spaced hyphen, … → … | On |
| Remove zero-width and non-breaking characters | Deletes zero-width spaces, joiners, soft hyphens and byte-order marks; converts non-breaking and other Unicode spaces to a normal space | On |
| Remove URLs | Removes http://, https:// and www. links |
Off |
| Remove emojis | Removes emoji, skin-tone modifiers and flag sequences | Off |
The options run in a fixed order (hidden characters, URLs, emojis, Markdown, punctuation, line joining, trimming, space collapsing, empty-line removal) so that each step sees clean input from the previous one.
Worked example
A paragraph copied from a PDF report:
The committee re-
viewed the proposal in
March 2025. Costs were
“within budget”.
Next steps are listed below.With Join broken lines switched on and the defaults left as they are, the output is:
The committee reviewed the proposal in March 2025. Costs were "within budget".
Next steps are listed below.The hyphenated re-viewed is rejoined because the fragment after the break starts with a lowercase letter; a genuine hyphenated compound at a line end, such as long-\nTerm, is left alone. The blank line survives because only single breaks are joined. The stats panel shows how many characters were removed, and the cleaned text can be copied or downloaded as a .txt file.
When to use which switch
- Pasting an AI answer into an email or document: keep the defaults. Markdown symbols and smart quotes disappear, the structure stays.
- Pasting into code, a CSV or a form field: switch on Remove empty lines too, and always keep the hidden-character option on. A single zero-width space in an email address or product code causes silent lookup failures.
- Quoting from a PDF or scanned document: turn on Join broken lines. Check headings afterwards; a heading directly followed by body text without a blank line will be joined to it.
- Preparing text for a translation tool or plagiarism checker: remove URLs and emojis so the tool counts only real words.
- Subtitles or social captions: turn markdown removal off if you use asterisks deliberately, and keep emojis.
Practical tips
- Text that looks fine but fails a search or comparison almost always contains a non-breaking space or a curly apostrophe. Run it through with only the two punctuation and hidden-character options ticked.
- Em dashes are converted to a spaced hyphen because that is the ASCII convention most style guides accept; if you need to keep them, untick the smart-punctuation option.
- The tool never changes letters or words, only spacing, symbols and punctuation, so it is safe to run on legal or academic text before checking it with the Grammar Checker.
- For long documents, clean section by section and read the joins; automatic line-joining cannot tell a list item from a wrapped line if the list has no bullets.
Frequently asked questions
What are hidden characters?
Zero-width spaces, non-breaking spaces and similar invisible Unicode characters that come along when copying from web pages or AI chats. They break search, forms and code. The cleaner removes them.