Remove Asterisks from AI Text

Why AI text is full of asterisks

ChatGPT, Claude and Gemini all write in Markdown. In a chat window that renders it, **important** shows up as bold. Copy the same text into a plain textarea — a CMS field, an email, a LinkedIn post, a Word document — and the markup arrives literally, asterisks and all.

This tool removes the markers while keeping the words. Bold and italic wrappers are unwrapped, ## headings become plain lines, and bullet characters like • and * become ordinary hyphens.

The invisible characters matter too

AI output often carries characters you cannot see: non-breaking spaces, zero-width joiners, and curly quotes. These break search-and-replace, corrupt CSV imports, and occasionally show up as question marks in older systems. They are stripped here as well.

Em dashes are a special case. Heavy em-dash use is one of the more recognisable habits of AI-written prose, so there is an option to convert them into commas — off by default, since plenty of human writers use them deliberately.

What is deliberately left alone

Asterisks that are not paired around text survive, so 2 * 3 = 6 is untouched. Underscores are only unwrapped when they sit at word boundaries, which means snake_case variables, file_name.txt and URLs containing underscores come through intact.

Everything runs in your browser. Text you paste here is never uploaded, which matters if the draft is confidential.