Clickbait Filter🔒 100% local
Analyze headlines for clickbait patterns. Scores 0-100% based on common clickbait triggers.
Related Tools
Scores a headline against 21 weighted word patterns - "you won't believe", "shocking", "doctors hate", "7 things" and similar - adding each match to a running total that is capped at 100 percent. The total is compared with the sensitivity slider, and every pattern that fired is listed by its raw pattern text so you can see which phrase was responsible. The headline is matched in the page; it is not sent anywhere for analysis.
How to use it
- Paste the headline into the box.
- Move the sensitivity slider if 50 percent is too strict or too loose for what you are checking.
- Press Analyze: the output shows the score, the threshold, the classification and the triggered patterns.
Worked example
You won't believe what happened next!becomesClickbait Score: 60.0%
Threshold: 50%
Classification: 🚨 CLICKBAIT
Triggered Patterns:
- you won't believe
- what happened nextTwo patterns match, 0.3 each, giving 60 percent. Scores are capped at 1.0, so a headline with many triggers tops out instead of exceeding 100 percent.
City council approves new bike lanebecomesClickbait Score: 0.0%
Threshold: 50%
Classification: ✅ Likely Genuine
No clickbait patterns detected.Nothing in the pattern list matched, so the output ends with the no-patterns line rather than a list.
When to use it
- Checking your own draft titles before they go into a feed, a newsletter subject line or an A/B test.
- Teaching a content team what the standard bait phrases are, with the weights visible.
- Triaging a batch of suggested headlines in a CMS and deciding which ones to rewrite first.
What it does not do
- It only knows the 21 phrases in its list, so a manipulative headline written in sober language scores 0 percent.
- Legitimate wire copy uses "breaking" and "urgent" honestly, so expect false positives on real news.
- The score is a sum of weights, not a calibrated probability - treat the threshold as a dial you tune per feed.
- One headline per run; there is no bulk or CSV mode.