Posting guidelines for the Text Humanizer Forum

These guidelines keep the Text Humanizer Forum useful for people who want to edit AI-assisted drafts so they read naturally. They are short on purpose. If a situation is not covered, ask what a careful editor would do, and do that.

Asking a question

• Paste a short sample, not the whole thing. A paragraph is plenty. It is enough to show the problem, and it keeps the thread readable. Do not paste a whole assignment, thesis chapter or confidential document.
• Say who the reader is. A cover letter for a hiring manager needs different edits from a blog post or an essay for a seminar.
• Say what you already changed. Tell us what the draft looked like before and what you edited by hand. This stops people from suggesting what you have already tried.
• Name the rule you have to meet. If a school, journal or employer has a policy on AI use, say what it is, so answers can fit it.
• Pick the closest section. A question about a flagged essay belongs under detector scores or academic writing, not workflow.

Writing an answer

• Show the edit. A before and after of one sentence teaches more than a paragraph of advice.
• Say why, in terms a reader would notice: sentence length, repeated openings, vague claims, stock transitions.
• Name the source when you rely on a policy, a study or a tool's documentation, and say when a rule differs between institutions.
• Say when you are not sure. A plain "I do not know" is better than a confident guess.
• Disagree with the text, not the person. Critique the sentence, the claim or the argument. Do not comment on who wrote it.

What we do not allow

• Requests to evade detectors. Do not ask how to make a text beat a particular detector or how to hide that AI was used where a rule requires you to say so. You can ask what a score means, why a false positive happened, or how to show that your own writing is your own.
• Requests to write graded work for someone. Do not ask other members to write, rewrite or complete an assignment, exam answer or application that someone else will be graded or judged on as their own work.
• Other people's work without permission. Do not post someone else's essay, article, message or code unless they have said you may. Quote briefly and say where it came from.
• Personal data. Do not post names, student numbers, email addresses, phone numbers, workplace details or anything else that identifies you or another person. Remove them from samples before pasting.
• Promotion and spam. No advertising, no link dropping, and no repeated posts of the same question.
• Promises about outcomes. Nobody here can promise that a text will pass a detector or be accepted by an instructor. Do not claim otherwise.

Moderation

Staff edit or remove posts that break these guidelines. Usually we edit first, for example by removing personal data from a sample, and we leave a note saying what changed. We remove posts that ask for evasion or graded work, that share another person's work without permission, or that are spam. Repeated breaches can lead to a posting restriction or removal of an account. If you think we got it wrong, email hello@texthumanizer.com with the thread address and we will take a second look.

Accepted answers

In question sections, the person who asked can mark one answer as accepted. Mark it when it solved your problem, or when it came closest. An accepted answer is shown first, and it tells readers that it worked for someone. It does not mean the answer is correct for every school, journal or employer. If a better answer arrives later, you can change your choice. In the two discussion sections, editing workflow and draft feedback, there is no accepted answer, because the point is conversation and comparing methods.

Links

You may link to sources, policies, tools and your own work when it helps answer the question. Say what the link is and why it is relevant. Links that exist only to promote a product or a site will be removed.

Reporting a problem

Use the report link on a post, or write to hello@texthumanizer.com with the thread address. Please include what is wrong and which guideline you think it breaks.