Editorial policy

How an article is written

Each article starts from a question readers actually ask, such as how long a cover letter should be or what to put in a follow-up email after an interview. A language model writes the first draft. We use AI for drafting because it is quick at laying out a structure and producing plain sentences. It is not a source, and nothing it writes counts as true until someone has checked it.

The draft then goes to the editor whose name appears on the byline. That edit pass is where the article takes its final shape. The editor cuts padding and generic advice, swaps a vague tip for a concrete example, rewrites sentences that read as if a machine produced them and checks every fact, as the next section describes. Nothing is published before that pass. So if AI helped with a sentence you are reading in one of our guides, a named editor has read it and answers for it.

How facts are checked

Every number, rule and claim about a tool, a platform or an institution is checked against a source before the article goes out. For a word limit on an application, that source is the form or the page that sets the limit. For a university rule, it is the university's own guidelines. For anything about Textlume, it is how the tool works today.

Language models can state something false in a confident tone. A fact that came from the draft therefore counts as unchecked until the editor finds it in a source. If no source confirms it, the sentence comes out. We don't keep a claim because it sounds right.

Figures about our own tools come from our own tests. The accuracy page, for example, reports how often the AI detector flagged human writing, with the sample size and the test date next to each figure, so you can judge how much weight it carries. When an article quotes one of those figures, it uses the same number and does not round it up into a stronger claim.

Who is on the byline

Articles appear under one of four editors, each with an area of their own. Carmen Echeverría, our academic writing editor, signs the guides on university writing. Jakob Wendt, our machine learning engineer, signs the pieces on how AI detection and the rewriting tools work. Ece Karaman, our education lead, signs the guides for students and teachers in Turkey. Malcolm Thorne, our workplace writing editor, signs the guides on writing for work, from cover letters to the email you send after an interview.

The name on an article tells you who did the edit pass and who answers for what it says. The accuracy page and our description of how we test the detector carry no personal byline. They report measurements from our own tests, and the company is responsible for them as a whole.

Corrections

Found a fact that is wrong, or a rule that has changed since we wrote about it? Write to support@textlume.com and tell us which article and which sentence. Dave from our support team replies within 24 hours. Once the error is confirmed, the editor corrects the article and the date on it changes to the day of the fix. A mistake in AI-drafted text is handled like any other mistake. If it got past the edit pass, it is ours to correct.

No paid placements

No company pays to be mentioned, linked or ranked in our articles. When a guide names another product, it is there because it is relevant to the question, and we describe it with facts its maker publishes. When a guide mentions a Textlume tool, it is there because it helps with that task, such as counting the words in a cover letter, and the guide also says what the tool can't do.

No article promises that a text will pass an AI detector, and none explains how to get around a school's or an employer's rules on AI. Where those rules matter, the article tells you to read them and follow them. Whether AI help is allowed on a particular assignment or application is decided by the people who set it, not by us.