Paste text that has arrived from a PDF, a spreadsheet, an email thread or a word processor, tick the fixes you need, and the box returns something you can use. Nine clean-up steps, plus find and replace, applied in one pass.
The page opens with four steps already ticked: trim the start of each line, strip trailing spaces, collapse repeated spaces, and remove blank lines. On the loaded sample that turns six lines into four and reports 8 characters removed.
The nine steps
| Step | What it removes |
|---|---|
| Line numbers | A number of up to three digits, then a full stop or bracket, then a space, at the start of a line |
| Bullets | A leading dash, bullet, middot, square, circle, asterisk or plus, then a space |
| Trim lines | Spaces and tabs at the start of every line |
| Strip trailing spaces | Spaces and tabs at the end of every line |
| Collapse spaces | Any run of two or more spaces or tabs, replaced with one space |
| Straighten quotes | Curly single and double quotes, replaced with straight ones |
| Normalise dashes | En, em and horizontal-bar dashes become a hyphen; ellipsis becomes three dots |
| Fix punctuation spacing | A space before , . ! ? ; :, and a missing space after them |
| Remove blank lines | Every line that is empty or only whitespace |
Line numbers and bullets are the two that save the most time on a paste from a PDF. 1. First item, 2) Second item and 3. Third all lose their prefixes, and a run of •, -, * and · lines all lose their markers.
Bullets is the one to check before you trust it. A line that begins with a hyphen is stripped whether it is a list item or a dash of dialogue, and a negative number at the start of a line loses its sign.
The order is fixed
Steps run from the top of the list down — the order in the table above — and find and replace runs last. Ticking them in a different order changes nothing, which matters for two pairs in particular.
Collapse spaces runs before fix punctuation spacing, so Hello , world first becomes Hello , world and then Hello, world. Run them the other way round and the space before the comma would survive.
Remove blank lines runs near the end, so it cleans up the gaps left by the earlier steps as well as the ones that were there already. Only steps that actually change something are listed in the report, which is why the default run on the sample says three steps applied rather than four — trimming the start of lines found nothing to do.
What fix punctuation gets wrong
This step is the one to use with care. To add a missing space after a full stop it looks for a full stop followed by a letter or digit, and inserts a space between them. A decimal point is a full stop followed by a digit, so numbers get broken:
3.14 and 2.5
Output:
3. 14 and 2. 5
Input:
Version 1.2.3 ships
Output:
Version 1. 2. 3 shipsLeave that step off any text containing decimals, version numbers, IP addresses or times. The other three punctuation cases it handles are a space before a comma or full stop, which it removes, and a missing space after a comma, which it adds.
It does not touch quotations. Hello , world ” becomes Hello, world ” — the space before the closing quote stays, because the rule only looks at the six punctuation marks and not at quotes or brackets. Two further gaps: it does nothing to an ellipsis, and repeated punctuation such as !!! is left alone.
Quotes and dashes
Straighten quotes covers the curly single and double forms, including the low and reversed variants that come out of Word and some PDFs, plus prime marks. Everything becomes a plain ' or ". That is what you want for code, CSV fields and plain-text email; it is not what you want for a page that should keep typographic quotes.
Normalise dashes turns en dashes, em dashes and horizontal bars into a hyphen, and a single ellipsis character into three dots. It does not add spaces around the dash, so word—word becomes word-word. If your style needs spaced dashes, do that with a find and replace afterwards.
Find and replace
The two fields at the bottom work as a literal search, not a pattern. Every occurrence of the exact string is replaced, and the report tells you how many matches there were. There are no wildcards, no case-insensitive option and no way to replace only the first match or only whole words. Searching for cat also changes the middle of category.
It runs after all the ticked steps, so it sees the cleaned text. That is usually what you want — replacing a curly quote after the quotes have been straightened would find nothing — but it does mean a replacement can be altered by an earlier step if you tick one that touches it.
Empty results and applying
If none of the selected steps match anything, the status says "Nothing matched" and the output is the input, unchanged. That is not an error: it usually means the text was already clean, or the step you expected is not the one that handles your problem.
Two things the tool does not touch: a non-breaking space between words survives collapse spaces, because that step only matches ordinary spaces and tabs; and tabs inside a line become a single space rather than a tab stop.
Nothing is written back to the input until you press Apply, so you can compare the two before committing. After that the original is gone — copy it first if you may need it. Everything runs in the tab and nothing is uploaded.