What it does
Catch the slop.
Sound like yourself.
Anti-Slop scores text the way a sharp editor would, then fixes what a find-and-replace can't: the rhythm, the buzzwords, the tells.
01 · Score
Instant 0–100 score
Select text and get a 0–100 human-read score and a verdict — reads human, some tells, or likely AI — in under a second, offline.
02 · Flag
Every tell, flagged
See the exact banned phrases and buzzwords, plus the structural signals: three-clause rhythm, uniform sentences, em-dash splices, canned conclusions.
03 · Fix
One-click humanize
The Pro AI rewrite breaks the robotic rhythm and rewrites in a natural voice, preserving your meaning, facts, numbers, and tone. Edit the result before you use it.
04 · Private
Free & offline
Detection and quick-clean run entirely in your browser. No account, no upload, no cost.
05 · Anywhere
Works everywhere
Gmail, Google Docs, your CMS, any draft or listing. Right-click a selection or paste into the popup.
06 · Crafted
Clean, fast, dark-mode
A crafted panel that respects your theme and stays out of the way. No clutter, no bloat.
Under the hood
How does Anti-Slop detect AI writing?
Anti-Slop scores text on a 0–100 human-read scale. Alongside a dictionary of banned phrases and buzzwords, it checks five structural signals that AI writing tends to have. Each signal that fires is shown, so you see exactly why the score landed where it did.
| Structural tell | What it looks like |
| Three-clause rhythm | Repetitive triads: “fast, clean, and simple.” |
| Uniform sentence length | Every sentence roughly the same length, no variation. |
| Decorative em-dashes | Em-dashes used for flourish instead of real punctuation. |
| Canned conclusions | “In conclusion…”, “At the end of the day…” |
| Hedge stacking | “It’s worth noting that it’s important to…” |
Banned phrases like “delve into,” “in today’s digital landscape,” and “leverage” are flagged on top of these signals. Detection runs offline in your browser — no account, no upload.
From selection to clean text in four steps
1
Select text on any page.
2
Right-click → Anti-Slop, or click the toolbar icon.
3
Read your score and the flagged phrases.
4
Quick-clean free, or AI-rewrite for a full pass.
Pricing
Free to detect.
Pro to rewrite.
Scoring and quick-clean are free forever. Upgrade only if you want the full AI rewrite.
Free
$0
Everything you need to detect AI writing.
- Unlimited scoring & flagging
- Deterministic quick-clean
- Runs 100% offline
- No account required
Add to Chrome
Pro
$19.99/mo
For heavy writers and editors.
- Everything in Free
- 1,000 AI rewrites / day
- Full humanizing pass
- Priority processing
Go Pro
Basic
$9.99/mo
For occasional rewriting.
- Everything in Free
- 100 AI rewrites / day
- Full humanizing pass
Get Basic
FAQ
Questions
Is Anti-Slop free?
Yes. Detecting AI writing and the deterministic quick-clean fixes are completely free and run entirely in your browser, with no account and no upload. The AI rewrite is the paid Pro feature: Basic is $9.99/month for 100 rewrites a day, Pro is $19.99/month for 1,000 a day.
How does it detect AI-written text?
It scores text on a 0–100 human-read scale. It looks for banned phrases and buzzwords, plus structural signals AI writing tends to have: the repetitive three-clause rhythm, uniform sentence length, decorative em-dashes, canned conclusions, and hedge stacking. It shows you exactly which phrases and signals fired.
Does my text leave my computer?
For the free features, no. The score, the flagged phrases, and quick clean all run locally in your browser. Only the Pro AI rewrite sends the selected text to our server to be rewritten, which is disclosed in our privacy policy.
Will the rewrite change my meaning?
No. The AI rewrite preserves your meaning, facts, numbers, and voice. It removes the AI tells and robotic rhythm without inventing claims or changing what you said. You can edit the result before using it.
Can it beat AI detectors?
Anti-Slop is a diagnostic and editing tool, not a way to fool detectors. It tells you which AI tells your text has and helps you write more like a human. We don't claim any output is undetectable.