AI text cleaner: remove ChatGPT formatting, citations and hidden characters
Copying from an AI chat brings along ### headings, **asterisks**, citation tokens and invisible Unicode characters. Clean them in one click and keep your words exactly as they are.
Ready when you are.
What it cleans
- Markdown syntax: headings (#), bold and italics (** and _), list markers, quotes, tables and code fences — the words stay.
- Citation markers from AI answers, such as 【4†source】, [oaicite:3] and the hidden cite tokens ChatGPT inserts around web results.
- Invisible characters: zero-width spaces and joiners, soft hyphens, byte-order marks and bidirectional controls.
- Extra blank lines and trailing spaces, without touching code blocks.
- Optional: curly quotes, long dashes and ellipses converted to plain characters.
What it does not do
- It does not rewrite, paraphrase or “humanize” your text.
- It cannot remove statistical watermarks and does not make text “undetectable”. Treat any tool that promises that with suspicion.
- It does not decide whether a text was written by AI.
Why does text copied from ChatGPT contain ** and ###?
Assistants write in Markdown. Many editors (e-mail, Word, social networks) do not render it, so the symbols appear literally. The cleaner removes the syntax and keeps the structure as plain text.
What are the strange characters like 【4†source】?
They are citation markers from web browsing or file search. They are meaningless outside the chat, so the cleaner removes them.
Does it remove hidden watermarks?
It removes invisible Unicode characters, which some people call “hidden watermarks”. Real AI watermarking is statistical (word choice) and cannot be removed by deleting characters, so we do not claim that.
Is my text stored?
No. The cleaning runs in JavaScript on this page; nothing is sent to a server or an AI model.
Need to clean whole conversations?
AI Vault imports your ChatGPT, Claude and Gemini exports, finds personal data across every message and exports exactly what you choose.
Open AI Vault