Changelog

Ghost 5.2 Mini is now the default lightweight humanizer model, built for fast, natural rewrites across everyday writing.

Ghost 5.2 Pro is tuned for more coherent structure, smoother flow, and better readability than the Mini model.

The humanizer model selector now includes the Ghost 5.2 generation while keeping previous-generation Ghost 5.1 and Ghost 4.6 options available under Legacy.

Humanized results now show the correct word count.

Sentence highlights now stay better aligned with your text.

Humanizer and scanner results now load more smoothly.

We improved the 5.1 humanizer server capacity so more users can process text at the same time with smoother availability during busy periods.

We refreshed our pricing so every plan shows the exact daily limits up front. Starter includes 50 humanizations/day, Plus includes 150/day, Pro includes 350/day, Business includes 800/day, and Scale includes 2,000/day. Paid plans support up to 5,000 words per input.

When you reach a daily usage limit, upgrade messages now show concrete plan limits instead of multiplier-only copy, making it easier to choose the right plan.

The 5.1 humanizer now supports major world languages — including German, Spanish, Portuguese, French, Italian, Dutch, Chinese, Japanese, and more. Paste text in your language and get natural, undetectable output without switching to English first.

The dashboard Support button now opens Intercom live chat immediately, with your account details attached automatically so our team can respond faster.

Resolved an edge case where longer inputs could be truncated before a sentence was fully rewritten. Your output now consistently finishes every sentence it starts.

Fixed an issue where some sentences weren't being highlighted after humanization. Every rewritten sentence is now properly mapped and visible in the UI.

We tuned the 5.1 humanizer to produce cleaner, more natural prose — fewer awkward phrasings, better flow, and less friction for your readers.

Our AI detector has been upgraded to V2 — significantly more accurate at identifying AI-written text and more consistent across different writing styles and lengths. Expect fewer false positives and sharper confidence scores.