The fastest way to ruin a humanized draft is to let the rewriter "improve" the words that were never supposed to change. A product name gets a synonym. A customer quote is paraphrased. A compliance line is softened. A figure is rounded.
ClearPen's protected text feature exists to stop that.
What Is Protected Text?
Protected text is an exact phrase in your draft that ClearPen is told not to change. A protected snippet is carried through the entire pipeline — the rewriting stages, the targeted paraphrasing, the grammar repair and the final polish — and is never altered along the way. The words you lock in are exactly the words that come out.
How to Protect Text in ClearPen
- Paste your draft into the input pane.
- Select the exact words you want to keep.
- Choose Protect. The snippet appears as a chip under the input so you can see everything that is locked.
- Run the humanizer as normal.
You can also protect a sentence from the AI-score heatmap: click it and choose Protect, and it will be kept as-is on the next run.
Each protected snippet must match text that appears exactly in your draft. ClearPen checks this on the server, so a typo in the selection simply will not lock anything.
What You Should Protect
| Content type | Examples | Why it must stay exact |
|---|---|---|
| Brand and product names | Product names, feature names, trademarked terms | Consistency and legal correctness. |
| Taglines and approved copy | Campaign lines, legal-approved claims | Already signed off; any change needs re-approval. |
| Direct quotations | Customer quotes, interview excerpts, testimonials | A paraphrased quote is no longer a quote. |
| Figures and data | Prices, percentages, dates, statistics | A rewritten number is a wrong number. |
| Legal and compliance wording | Disclaimers, terms, regulated phrasing | Meaning depends on the precise words. |
| Technical terms | API names, code identifiers, model numbers | Synonyms break technical meaning. |
| References | Reference list entries, source titles | Must match the source exactly. |
Protect Less, Get Better Results
Every word you lock is a word ClearPen cannot rewrite — so if you protect a long passage that reads as AI, it will likely still read as AI afterwards. The best results come from protecting the shortest exact phrase that has to stay:
- Protect "Acme CloudSync Pro", not the whole sentence that mentions it.
- Protect the figure "37.4%", not the paragraph explaining it.
- Protect the quote itself, not the sentence introducing it.
That way ClearPen can still rewrite everything around the locked words. If ClearPen ever returns your original text because the rewrite scored worse, the notice will suggest exactly this: protect only the phrases that must stay, rather than whole sentences.
Limits
- Up to 10 protected snippets per run.
- Each snippet up to 400 characters.
- Each must appear exactly in the text you are humanizing.
How Protected Text Fits With ClearPen's Other Controls
Protected text pairs naturally with ClearPen's other precision tools:
- Spot Fix rewrites one chosen sentence after a run — the inverse of protection.
- Strength levels (Gentle, Balanced, Max) set how much of the unprotected text changes. See how ClearPen works.
- US or UK spelling sets the convention for the final polish, so your spelling stays consistent with your brand style.
Together they give you control at the level that matters for professional writing: this word stays, this sentence changes, everything else reads naturally.
Humanize a draft without losing a single brand term
Lock your names, quotes and figures, then compare the before and after scores. Free plan, no credit card.
Related Reading
Frequently asked questions
- Can I stop an AI humanizer from changing certain words?
- In ClearPen, yes. Select the exact text in your draft and choose Protect. Protected snippets are kept word for word through every stage of the rewrite.
- How many phrases can I protect in ClearPen?
- Up to 10 protected snippets per run, each up to 400 characters long. Each snippet must match text that appears exactly in your draft.
- What should I protect when humanizing text?
- Anything that must stay exact: brand and product names, taglines, direct quotations, statistics and figures, legal or compliance wording, technical terms, code identifiers and reference entries.
- Does protecting text affect the AI score?
- The more text you lock, the less ClearPen can rewrite, so a long protected passage that reads as AI may still score that way afterwards. Protect the shortest exact phrase that must stay — a name or a figure — rather than a whole sentence, so ClearPen can still rewrite the words around it.
- Does ClearPen keep headings and structure?
- Yes. Headings pass through the pipeline unchanged and the document is reassembled in its original structure, whether or not you protect anything.