Amazon without the knockoffs
Purifying Your Amazon Experience with Knockoff
Knockoff is designed to transform Amazon into a marketplace where only brands with an actual reputation to protect remain. It systematically scrubs "trademark-squatting" pseudo-brands—those nonsensical names like
SZHLUXorHORUSDY—from your search results.
Get Started: [Add to Chrome] | [Star on GitHub] License: Free, Fair Source (FSL) No account required Zero tracking.
The Problem: Search Result Pollution
When you search for a common item, like a dog bed, Amazon often floods the page with generic, randomly named entities.
Example Search: amazon.com/s?k=dog+bed
| Brand | Product Description | Status | Rating | Price |
|---|---|---|---|---|
| XL Extra Large Dog Bed, Washable | Pseudo-brand | ★★★★☆ | $35.99 | |
| Casper | Plush Memory Foam, Large, Sand | Established | ★★★★☆ | $224.10 |
| Orthopedic Dog Beds for Extra Large | Pseudo-brand | ★★★★☆ | $39.99 | |
| Carhartt | Durable Canvas Pet Bed, Water Resistant | Established | ★★★★☆ | $139.99 |
Result: 25 items filtered. Junk is either flagged, dimmed, or completely removed.
How the Engine Works
The entire appraisal process happens locally within your browser every time a page loads. The logic follows this flow:
The Technical Heuristics
To identify unknown brands, Knockoff analyzes the "linguistic signature" of the name. We can represent the probability of a brand being a "pseudo-brand" () based on features like:
Specifically, it looks for:
ALL-CAPSstrings of a specific "squat-length."- A noticeable absence of vowels.
- Unpronounceable sequences of consonants.
Customization & Control
You have total authority over how your results are displayed.
- Adjust Strictness: Choose between Relaxed, Standard, or Strict modes.
- Visual Filtering:
- Hide: Listings vanish (though a one-click reveal is available).
- Dim: Items fade into the background until you hover over them.
- Label: A warning chip is attached to the listing.
- Ad Removal: A single toggle allows you to strip out all
Sponsoredlistings. - Personal Lists: You can manually trust or block any specific brand.
Community-Driven Accuracy
Because the brand lists are simple plain-text files, the community can maintain them with extreme efficiency.
Reporting a mistake is simple:
- Use the one-click report tool inside the extension.
- A curated list is updated.
- The fix reaches all users within 24 hours without needing a browser update.
For developers, fixing a misclassification is as simple as a one-line Pull Request:
- EHEYCIGA
+ GenuineBrandName
Whether you are adding new brands, refining the heuristics, or reporting errors, the system is built to be transparent and collaborative.