AI Slop Detector
AI Slop Detector is Jon Kraayenbrink's (@kraayenJon) free scanner linked from his September 2026 launch post.
by @kraayenJon
Overview
AI Slop Detector is Jon Kraayenbrink's (@kraayenJon) free URL scanner. A September 2026 clip shows Jev checking a site for thirty five parallel tells of AI slop, including purple gradients, emoji headers, seamlessly phrasing, fake testimonials, and bento grids, in about two hundred forty three milliseconds for about $0.00015 of tokens. Headless Chrome captures the page; vision describes visuals; Jev answers parallel Noul questions plus a whole-page score; DeepSeek writes the verdict text from found tells. The live tool allows two scans per day without signup on the public free-tools URL linked in the post.
Problem: Buyers and readers need a fast signal that a landing page was assembled from generic AI layout patterns, not a careful human edit.
Built for: Marketers, founders, and designers auditing competitor sites or their own drafts before launch.
First indexed on Jev Directory: 2026-09-21
Creator and team
- Name
- Jon Kraayenbrink
- Handle
- @kraayenJon
“In 243 ms it checked a website for 35 tells of ai slop. Paste any url, get a slop score.”
How Jev is used
- Role in the product flow
- Parallel Noul tells plus page-level Score after vision description
- Primitives
- NoulScore
- State in
- Rendered page screenshot or vision-derived description from headless Chrome (per launch clip).
- Decision out
- Per-tell Noul flags, aggregate slop score, and DeepSeek-generated verdict copy listing matched patterns.
- Fetch and render target URL in headless Chrome
- Vision model describes layout and copy cues
- Jev runs about thirty five parallel Noul questions on tells
- Whole-page Score summarizes slop level
- DeepSeek writes human-readable verdict from hits
The slop detector keeps expensive vision and cheap judgment separate. Chrome plus a vision description produces structured state; Jev fires dozens of tiny Noul questions in parallel so each tell is typed instead of buried in one chat essay. A Score primitive rolls the tells into a single slop number suitable for sorting URLs. DeepSeek only narrates the evidence Jev already flagged, which keeps latency sub-second in the attributed clip. Operators paste any public URL into the free tool linked from Kraayenbrink's post. This profile cites the X launch and live scanner, not third-party directories.
Sourced performance claims
- About two hundred forty three milliseconds and about $0.00015 of tokens for thirty five tells in the launch clip.Source: kraayenJon X post 2101157548346794059
- Free tier: two scans per day without signup (per launch post).Source: kraayenJon X post linked free tool
Features and stack
Features
- Thirty five parallel slop tells per URL
- Sub-second Jev pass in the public clip
- Free daily scans on the linked tool
- Showcase video embed from @kraayenJon
Stack
- Headless Chrome capture
- Vision description
- TypeSafe System One
- DeepSeek verdict copy
Pricing: Two free scans per day on the public tool; additional usage not documented in the clip.
Demo
Showcase clip with the same lightbox player as the homepage. Click to play.
AI Slop Detector for landing pages
Thirty five parallel Noul tells catch purple gradients and fake testimonials before your designer notices. Chrome renders; Jev judges; DeepSeek narrates the roast.
Links
FAQ
- Does Jev write the final paragraph?
- Jev answers the parallel tell questions and scores the page. DeepSeek generates the readable verdict from the tells Jev flagged, per the launch thread.
- What counts as a tell?
- The clip names patterns like purple gradients, emoji headers, seamlessly phrasing, fake testimonials, and bento grids among thirty five parallel checks.
- Is there an API?
- This listing documents the free URL tool linked in the X post. No public API was claimed in the September 2026 clip.
Related learn guides
Original Jev guidance that pairs with this product pattern.
- Jev use cases
The patterns builders actually search for: moderation, routing, triage, RAG verify, and agent gates.
- Jev vs LLM classification
When to gate with System One probabilities instead of asking a chat model to label things.
Related products
Hand-picked neighbors with rich profiles or overlapping tags.