Slop*Detector

SUSPICIOUSLY VIBEY — slop score 62/100

The evidence here points pretty clearly to a generated build rather than a hand-designed site. The gradient-washed hero with its rounded 'Premium features available' pill, the indigo/violet color scheme running through 60-plus utility classes, and the repeated 3-card grids that just restate the same six scripts under different headers (Learn Hiragana, Read Hiragana, Hiragana Table) all point the same direction. The card copy seals it: phrases like 'Comprehensive katakana learning with practice exercises and mnemonics' and 'Master hiragana through interactive exercises' follow the same adjective-plus-paired-noun template with no product-specific detail, and an em dash even shows up in a section header. Add the mismatched AI-illustration style across thumbnails, the stock 'Made with ❤️' footer, and the cross-promo box for a sister Korean-practice site built on the identical template, and you get a site assembled from generator defaults rather than edited by a person with a design point of view. A 62 feels right: nothing here is catastrophic on its own, but the pattern is consistent across visuals, layout, and copy, which is what pushes this into suspiciously vibey territory instead of borderline.

Findings

• Pill tags — Little fully-rounded pill badges sprinkled across the page ("New", "Beta", category chips) are a default garnish of vibe-coded UI.
• Gradient backgrounds / gradient text — Large gradient washes and gradient-clipped headline text (usually blue→purple) are the default decoration of generated UI.
• Indigo/violet accent palette — The blue-purple-violet accent range is the statistical center of mass of AI-generated color schemes.
• Untouched shadcn/Tailwind defaults — shadcn components, lucide icons, and Inter/Geist/Space-Grotesk font stacks left at their defaults mean the model picked the design, not a person. Instrument Serif as an accent font is a separate, independently corroborated tell.
• Glassmorphism — Frosted-glass backdrop-blur panels are a recurring generated-UI decoration.
• "Built with ❤️" footer — The affectionate footer sign-off ships with every generated template.
• Feature-tile grid — Multiple 3-card grids repeat identical icon+kanji+subheader+blurb structure across several sections ('Learn Characters', 'Reading Practice', 'Character Tables', 'Other Tools'), a hallmark of generated section templating.
• Announcement pill above hero — A rounded capsule sits directly above the H1, the canonical generated-hero opener pattern.
• Indigo/violet accent range — The primary buttons, top banner, and links all use the same indigo/violet hue, the statistical default AI accent color, with no alternate palette decision visible.
• Template skeleton / redundant sections — The page stacks near-identical sections (Learn Characters, Reading Practice, Character Tables, Other Tools) that largely restate the same six scripts in slightly different framings, suggesting slots filled by a generator rather than an edited information architecture.
• AI-rendered illustration with inconsistent/garbled content — The character thumbnails feature a repeating anime-style mascot in oversaturated, semi-photoreal digital-painting style, with background elements resembling almost-text glyphs and inconsistent scene compositions (classroom, neon city, sparkle background) that read as separate AI image-generation prompts rather than a coherent illustration set.
• Inconsistent illustration style across page — Thumbnail art shifts between semi-realistic anime shading, flat neon-arcade coloring, and moody desaturated palettes across cards in the same grid, indicating no unified art direction.
• Sister-site cross-promo box mirroring template pattern — The 'Also Check Out' box promoting a 'Practice Korean' sister site with generic copy and a CTA button matches the same indigo-gradient, stock-CTA aesthetic seen elsewhere, suggesting a duplicated template applied across multiple site instances.
• Generic educational disclaimer boilerplate — The footer disclaimer reads as stock legal filler generated to cover the bases rather than a specific, considered policy statement.
• Em dashes — Em dashes are the single most reliable punctuation tell of LLM copy. Human marketing copy almost never uses them; models reach for them constantly.
• AI vocabulary — Words statistically overrepresented in generated text: "seamless", "leverage", "unlock", "empower", "elevate", "robust" and friends.
• Em dash usage — An em dash appears in a section header, which the style guide flags as a banned punctuation habit associated with generated text.
• Templated, interchangeable card descriptions — Nearly every feature card follows the identical generative pattern of adjective + noun + 'with/through' + two paired nouns, producing descriptions that could be swapped between products with minimal edits — a hallmark of AI-filled boilerplate rather than copy written with knowledge of the specific tool.
• Repetitive lexical crutches — Words like 'Comprehensive,' 'Systematic,' and 'Master' recur across nearly every card description instead of varied, feature-specific language, suggesting a single generation pass reusing a small vocabulary bank rather than a person describing each tool distinctly.
• Low specificity, altitude-of-features language — Descriptions stay at the generic 'features' altitude (exercises, references, practice) without any concrete numbers, counts, or unique product details (e.g., how many kanji, which JLPT levels, what the mnemonic system actually is), consistent with copy generated without real product knowledge.