Assistant output, invisible characters and your submission
What goes wrong
You used an assistant somewhere in your process, perhaps only to rephrase a paragraph or to check your grammar, and your institution runs submissions through a similarity or AI-detection service. What you cannot see in your file may be counted against you.
Copying and pasting carries characters along with the words. Text that has been through a chat interface routinely contains narrow no-break spaces or zero-width characters, and they survive into your word processor untouched. This is true even when the words are entirely yours and the assistant only reformatted them.
What to do about it
Know what is in your file before you submit it. Paste your draft into the cleaner here and look at the census, then decide what to do with what you find. Be honest with yourself about the second half too: cleaning characters does not change how the writing reads, and if your institution's rule is about the process rather than the file, no tool addresses that. This page will not help you misrepresent your work, and we would rather say so than sell you something on a false promise.
Questions
Does an invisible character prove I used an AI?
No. They come from copying and pasting generally, including from web pages and PDFs. But a detector may still count them, which is why it is worth knowing whether they are there.
Will this get me past my university's checker?
It is not built for that and we are not claiming it. It removes characters you did not choose. It does not change your writing, and it cannot change how you produced it.