Why does my writing get flagged as AI even though I wrote it?
AI detectors key on statistical patterns, not truth: uniform sentence length, high em-dash density, stock vocabulary like "delve" and "leverage", and constructions like "not just X, but Y". Human writing — especially formal, formulaic or non-native English — can contain these patterns naturally. Red Pencil shows you exactly which patterns appear in your text, so you remove them deliberately instead of guessing why a detector is angry.
What words make text sound like it was written by AI?
The 2026 consensus list is led by: delve, tapestry, leverage, harness, unlock, foster, bolster, pivotal, groundbreaking, transformative, seamless, robust, meticulous, testament, paradigm, elevate, showcase, plethora and myriad. Red Pencil checks ~60 of these words, 19 stock phrases and 7 structural patterns — and updates the list monthly as models change their habits.
How do I remove em dashes from AI text?
Paste your text above. When em-dash density passes the human baseline (about one per 140 words), every em-dash gets flagged with a one-click fix to a comma or period. Keep your best one — a single em-dash reads fine; five per paragraph reads machine-made.
Is Red Pencil an AI humanizer?
No — deliberately. Humanizers paraphrase your text through another AI, which is why their own reviews complain about mangled meaning and awkward output. Red Pencil is a linter: it marks each tell, explains why it reads as AI, and you approve every single change. Your sentences stay yours.
Does Red Pencil upload my text anywhere?
No. The whole rule engine is JavaScript running in your browser. Your text never leaves your device — which is also why there are no word limits, no daily credits and no account. Paste an NDA-covered client draft with a clear conscience.
Will this make my text pass GPTZero, Turnitin or Originality?
Honest answer: it removes the known pattern-level tells those tools and human readers key on, which typically lowers scores — but nobody can guarantee a detector result, and anyone who promises one is selling you something. Red Pencil's goal is writing that reads like you wrote it, because you did.
How much does Red Pencil cost?
Checking is free forever, with no word limit. Pro is $4.99 (₹399) — once — and unlocks Fix All, long-document mode and export. The optional Rewrite plan is $2.99/month (₹249), because AI rewriting is the one feature that costs us real compute per click; everything that runs in your browser stays one-time or free. We will never move the linter itself behind a subscription.
Will the AI rewrite pass Red Pencil's own check?
Yes — by design, and this is the difference from every humanizer. The rewriter is given Red Pencil's own rule set as its instructions, works one sentence at a time, and the moment a rewrite is inserted your text is automatically re-checked against the same rules. If anything slips through, it is flagged immediately and fixable in one click — nothing gets silently shipped. You approve or reject every rewrite; blind whole-document paraphrasing is exactly what we refuse to build.
How do the rules stay current as AI models change?
The rule set follows the July 2026 editorial consensus — including Wikipedia's "Signs of AI writing" project and working style guides — and ships as a versioned file updated monthly. Because Red Pencil is a website, every user gets the new rules instantly, with nothing to reinstall.
Does it work on essays, emails, LinkedIn posts and long documents?
Yes — anything you can paste. Free handles everyday drafts; Pro's long-document mode is tuned for 10,000+ word reports, theses and manuscripts, with Fix All applied in one pass.