Detect, humanize, paraphrase & attribute — one loop.

An explainable AI-content detector — burstiness, perplexity, cliché lexicon and punctuation tells — with a per-sentence heatmap and an honest inconclusive band.

◈  Detect AI-likeness  →

Rewrite then re-score in a loop — detector-targeted, style-imitating, meaning preserved.

✨ Deep Polish — re-attack paragraphs still reading AI 🧠 Auto (self-learning) — let the engine pick strength/passes from what actually beats the real detector ♁  Humanize  →

Reword in multiple modes for clarity and flow, with detector-aware optimisation and a change tracker.

Detector-aware optimisation ↻  Paraphrase  →

Attribute each passage as human, AI-assisted/mixed, or generative-AI — an origin distribution across the whole document.

⊚  Attribute authorship  →