How to clean text you pasted from the web

Text that looks clean can still contain characters you did not type. A paragraph copied from a web page may include a non-breaking space that refuses to wrap. A PDF may add a line break in the middle of a sentence and a hyphen that was only there for the column width. Email clients like curly quotes. A zero-width space can sit between two letters and break a search for a word you can see perfectly well.

What the cleaner removes

With the default steps on, Text Cleaner does six practical things:

• It composes Unicode to a stable form so the same accented letter does not exist as two different sequences.
• It deletes zero-width spaces, joiners used outside emoji, and the byte-order mark.
• It deletes control characters and leaves ordinary line breaks and tabs alone.
• It turns curly quotes and dashes into straight quotes and hyphens.
• It collapses repeated exclamation marks and question marks.
• It trims lines, collapses runs of spaces, and reduces stacks of blank lines.

Each step has a switch. Poetry and dialogue sometimes need the blank lines. Published prose sometimes needs the curly quotes. Turn a step off instead of cleaning twice and hoping.

What it does not do

The cleaner does not join words that a PDF split with a hyphen at the end of a line. It does not paraphrase, translate, or "improve" the wording. It does not send the text anywhere. If the paste is a table, clean it and then read a few cells, because aggressive space collapsing can make aligned columns look like sentences.

Duplicate lines are a separate problem. A list of keywords exported from a spreadsheet often repeats the header or the same query. The duplicate line remover keeps the first copy and the original order. Use it after the cleaner when the draft has both issues.

A short habit before you publish

Paste the source into the cleaner, copy the result into the CMS, and then run the word counter or the title checker on the text you are actually publishing. Counting the messy version wastes the edit. If you are signed in, you can save either version as a project. That save is explicit. Closing the tab without saving leaves the text on the device, and a refresh clears the editor.