Normalize pasted AI-generated text
Normalize punctuation, spacing quirks, and invisible characters from copied AI drafts — the entire cleanup runs in your browser.
Three quick steps
- Paste your text into the editor below.
- Fix & Copy to normalize the odd characters.
- Done — the cleaned text stays on your device.
AI drafts often add “smart punctuation,” narrow spaces, and zero-widthjoiners — all highlighted here.
Paste your text
We only run in your browser. No uploads, no logs.
How it works
- Detects the most common cross-model character issues: curly punctuation, zero-width joiners, invisible spaces, and ornamental bullets.
- One click replaces what can be safely normalised; anything subjective stays highlighted for manual review.
- Does not touch your original wording — only characters are changed.
- Runs locally in the browser so sensitive prompts remain private.
What this page specifically fixes
This is the concrete character-level scope for this cleaner. If your problem matches one of these rows, this page is the right place to start.
| Pattern | Handling | Why it matters |
|---|---|---|
| Cross-model punctuation and spacing issues | Normalize safe replacements | The page catches the characters that show up repeatedly across ChatGPT, Gemini, Claude, and similar tools. |
| Invisible separators and odd spaces | Flatten to regular spacing | These cause the biggest copy/paste surprises in ATS, CMS, and markdown workflows. |
| Decorative bullets and polished glyphs | Flag or simplify them | They often travel badly once generated text leaves a styled interface. |
Where these characters usually come from
- Any mixed AI workflow where you are unsure which model produced the draft.
- Teams pasting model outputs through docs, tickets, and CMS tools.
- Editing pipelines where you want a single broad cleanup pass before rewriting.
Before and after examples
Before
“Here’s the draft…” — with a hiddenspace
After
"Here's the draft..." - with a hiddenspace
A broad cleanup pass removes the most common copy/paste issues in one sweep.
Before
• Summary • Risks • Next steps
After
- Summary - Risks - Next steps
Decorative list markers become simpler and more portable.
Before
Plan updated
After
Plan updated
Odd spaces are flattened before the text enters stricter systems.
Best for this cleaner
Use this page as a pre-flight check before you publish, submit, or automate text. It is strongest at character-level cleanup, not rewriting your message.
- Quick cleanup when you are unsure which model generated the draft.
- Standardizing mixed-source text before passing to teammates.
- Reducing obvious typography tells before manual rewriting.
Manual review checklist
After fixing characters, do one quick human pass. This catches tone or meaning issues that no character normalizer can decide for you.
- Edit repetitive sentence patterns; character cleanup alone is not enough for originality.
- Keep domain-specific symbols and units only where they are semantically required.
- Validate final text in the exact platform where it will be published.
When to be careful
- This page fixes typography and character-level problems, not originality, truthfulness, or tone.
- If a document depends on exact symbols, verify those lines manually before publishing.
- Use this as the broad default page, then open a specific guide if one issue keeps recurring.
Mini FAQ
Does this rewrite AI-generated text for me?
No. This page only normalizes characters and spacing. You still need human review for structure, clarity, tone, and originality.
Can I undo the changes?
Yes. Click the Clear button to revert, or simply refresh the page — nothing is saved.
Does it support non-English text?
Yes. The character catalogue covers multiple scripts and punctuation systems. The current review build is English-only, but the cleaner still detects those characters.