TurnAuthentic scores any document against 30+ transparent writing signals, points out the exact sentences that read as AI-generated using Claude or OpenAI, and rewrites flagged text into a natural, human voice you can download in one click.
412 words · 19 sentences
No hidden model, no mystery score. Every signal in the report is explained, and every AI-assisted feature is opt-in.
Burstiness, lexical diversity, repeated phrasing, stock AI phrases, punctuation variety, sentence rhythm, and more — every signal shown with its own score and weight, not a single opaque number.
Opt in and Claude or OpenAI reads the document sentence-by-sentence, flagging exactly which lines read as AI-generated and why — layered on top of the heuristic report, never replacing it.
Pick a tone — casual, professional, persuasive, storytelling — and a rewrite intensity. The LLM rewrites the text to sound naturally human while keeping every fact intact, then falls back gracefully if the API is ever unavailable.
Bring a .txt, .docx, or .pdf, or just paste text directly. Download the rewritten version as a plain-text file when you're done, and revisit every past scan and rewrite from your history.
Drop in a .txt, .docx, or .pdf, or paste text straight in. Nothing leaves your account until you ask it to.
See a full signal breakdown by category, plus — if you opt in — the exact sentences an LLM flags as AI-like, with its reasoning.
Choose a tone and intensity, let the LLM rewrite the flagged text into a natural voice, then download the result as a .txt file.