Guides
Why AI text arrives broken, and how to fix it. Every guide ends with a tool that does the work in one click.
Spotting AI text
Seven signs text was written by AIThe seven formatting and style patterns that show up in AI-generated text, how to check for each one, and why none of them is proof on its own.Why does ChatGPT use so many em dashes?ChatGPT leans on em dashes because it learned from edited prose. Here is why the habit is so consistent, and how to remove them cleanly.Does ChatGPT hide invisible characters in your text?AI output really does contain invisible Unicode. Here is what is actually in there, where it comes from, and whether any of it is a watermark.
Fixing AI output
Why does pasted AI text show ** and ## in Word?Those asterisks and hash marks are Markdown. Here is why Word shows them literally, and the three ways to paste AI output with formatting intact.How to paste a ChatGPT table into Excel or Google SheetsAI tables paste into one column because they are plain text, not cells. Here is how to get real rows and columns in Excel, Sheets, or Numbers.How to export a ChatGPT conversation to PDF or WordPrint-to-PDF clips your chat and loses formatting. Here is how to turn a conversation into a clean, readable PDF or Word document instead.Why smart quotes break your code, JSON, and CSVCurly quotes are different Unicode characters from straight quotes. Here is why parsers reject them, and how to normalize text safely.
Reference
Em dash vs en dash vs hyphenThree characters, three jobs. When to use each dash, how to type them on any keyboard, and why swapping one for another breaks things.How many tokens is my prompt?Tokens are sub-word fragments, not words. Here is how tokenization works, why counts differ between models, and how to estimate cost.Invisible Unicode characters: what they are and what they breakA reference to zero-width spaces, soft hyphens, BOMs, and direction marks — what each one does, where it comes from, and how to remove it.