| 59.09% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 22 | | adverbTagCount | 4 | | adverbTags | | 0 | "Eva said carefully [carefully]" | | 1 | "Eva said sharply [sharply]" | | 2 | "Eva said quietly [quietly]" | | 3 | "Quinn said slowly [slowly]" |
| | dialogueSentences | 43 | | tagDensity | 0.512 | | leniency | 1 | | rawRatio | 0.182 | | effectiveRatio | 0.182 | |
| 70.21% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1175 | | totalAiIsmAdverbs | 7 | | found | | 0 | | | 1 | | | 2 | | | 3 | | | 4 | | adverb | "barely above a whisper" | | count | 1 |
|
| | highlights | | 0 | "slowly" | | 1 | "gently" | | 2 | "carefully" | | 3 | "sharply" | | 4 | "barely above a whisper" |
| |
| 100.00% | AI-ism character names | Target: 0 AI-default names (17 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 61.70% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1175 | | totalAiIsms | 9 | | found | | | highlights | | 0 | "echoing" | | 1 | "perfect" | | 2 | "pristine" | | 3 | "scanning" | | 4 | "etched" | | 5 | "unwavering" | | 6 | "flicker" | | 7 | "whisper" | | 8 | "flickered" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 0 | | maxInWindow | 0 | | found | (empty) | | highlights | (empty) | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 1 | | narrationSentences | 81 | | matches | | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 1 | | hedgeCount | 0 | | narrationSentences | 81 | | filterMatches | | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 101 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 33 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 1165 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 16 | | unquotedAttributions | 0 | | matches | (empty) | |
| 0.66% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 48 | | wordCount | 837 | | uniqueNames | 7 | | maxNameDensity | 2.99 | | worstName | "Quinn" | | maxWindowNameDensity | 4.5 | | worstWindowName | "Quinn" | | discoveredNames | | Detective | 1 | | Harlow | 1 | | Quinn | 25 | | Victorian | 1 | | Kowalski | 1 | | Eva | 15 | | Morris | 4 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Victorian" | | 3 | "Kowalski" | | 4 | "Eva" | | 5 | "Morris" |
| | places | (empty) | | globalScore | 0.007 | | windowScore | 0.167 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 62 | | glossingSentenceCount | 1 | | matches | | 0 | "looked like an expensive suit" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1165 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 101 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 42 | | mean | 27.74 | | std | 15.6 | | cv | 0.562 | | sampleLengths | | 0 | 60 | | 1 | 21 | | 2 | 69 | | 3 | 18 | | 4 | 19 | | 5 | 39 | | 6 | 60 | | 7 | 9 | | 8 | 20 | | 9 | 39 | | 10 | 19 | | 11 | 42 | | 12 | 16 | | 13 | 36 | | 14 | 8 | | 15 | 30 | | 16 | 18 | | 17 | 11 | | 18 | 5 | | 19 | 45 | | 20 | 3 | | 21 | 17 | | 22 | 34 | | 23 | 15 | | 24 | 8 | | 25 | 29 | | 26 | 37 | | 27 | 7 | | 28 | 34 | | 29 | 33 | | 30 | 18 | | 31 | 50 | | 32 | 31 | | 33 | 13 | | 34 | 36 | | 35 | 38 | | 36 | 27 | | 37 | 41 | | 38 | 25 | | 39 | 31 | | 40 | 41 | | 41 | 13 |
| |
| 92.27% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 3 | | totalSentences | 81 | | matches | | 0 | "been polished" | | 1 | "been sealed" | | 2 | "were etched" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 155 | | matches | (empty) | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 7 | | semicolonCount | 0 | | flaggedSentences | 7 | | totalSentences | 101 | | ratio | 0.069 | | matches | | 0 | "The smell hit her immediately—copper and ozone, with something else underneath." | | 1 | "Her leather satchel hung heavy at her hip, and she kept tucking escaped curls behind her left ear—a nervous tell Quinn had noticed during their previous encounters." | | 2 | "No footprints—not even his own." | | 3 | "His clothes were immaculate—not a wrinkle, not a speck of dust except where he touched the floor." | | 4 | "The needle didn't track north—it kept pointing down the tunnel, unwavering." | | 5 | "Eva's green eyes met hers, and for a moment, Quinn saw something flicker there—knowledge, fear, or maybe both." | | 6 | "For just a second, Quinn could have sworn she smelled it again—that same ozone scent from the night she'd lost Morris." |
| |
| 97.23% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 834 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 36 | | adverbRatio | 0.04316546762589928 | | lyAdverbCount | 15 | | lyAdverbRatio | 0.017985611510791366 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 101 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 101 | | mean | 11.53 | | std | 7.06 | | cv | 0.612 | | sampleLengths | | 0 | 26 | | 1 | 15 | | 2 | 11 | | 3 | 8 | | 4 | 19 | | 5 | 2 | | 6 | 11 | | 7 | 31 | | 8 | 27 | | 9 | 13 | | 10 | 5 | | 11 | 12 | | 12 | 7 | | 13 | 4 | | 14 | 15 | | 15 | 9 | | 16 | 11 | | 17 | 7 | | 18 | 20 | | 19 | 3 | | 20 | 5 | | 21 | 25 | | 22 | 9 | | 23 | 15 | | 24 | 5 | | 25 | 6 | | 26 | 3 | | 27 | 13 | | 28 | 8 | | 29 | 9 | | 30 | 7 | | 31 | 12 | | 32 | 11 | | 33 | 7 | | 34 | 7 | | 35 | 17 | | 36 | 16 | | 37 | 4 | | 38 | 18 | | 39 | 14 | | 40 | 6 | | 41 | 2 | | 42 | 19 | | 43 | 11 | | 44 | 8 | | 45 | 10 | | 46 | 8 | | 47 | 3 | | 48 | 3 | | 49 | 2 |
| |
| 74.59% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 3 | | diversityRatio | 0.46534653465346537 | | totalSentences | 101 | | uniqueOpeners | 47 | |
| 82.30% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 2 | | totalSentences | 81 | | matches | | 0 | "Just like another compass had" | | 1 | "Maybe hadn't even seen it" |
| | ratio | 0.025 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 14 | | totalSentences | 81 | | matches | | 0 | "She descended the rusted maintenance" | | 1 | "Her leather satchel hung heavy" | | 2 | "She approached slowly, pulling on" | | 3 | "She'd already noticed." | | 4 | "She'd checked them all herself" | | 5 | "His clothes were immaculate—not a" | | 6 | "She pointed to the victim's" | | 7 | "She paused, seeming to weigh" | | 8 | "She felt the hair on" | | 9 | "she nodded toward the compass" | | 10 | "She owed Morris that much." | | 11 | "she told Eva" | | 12 | "She fixed the younger woman" | | 13 | "she emphasized the word, \"—goes" |
| | ratio | 0.173 | |
| 64.94% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 64 | | totalSentences | 81 | | matches | | 0 | "The abandoned platform stretched before" | | 1 | "She descended the rusted maintenance" | | 2 | "The smell hit her immediately—copper" | | 3 | "Something that made her jaw" | | 4 | "Eva Kowalski called from below," | | 5 | "Quinn skipped it, landing with" | | 6 | "The redheaded woman stood near" | | 7 | "Her leather satchel hung heavy" | | 8 | "Quinn surveyed the scene, noting" | | 9 | "Eva's green eyes darted toward" | | 10 | "Quinn followed her gaze." | | 11 | "The body lay twenty feet" | | 12 | "She approached slowly, pulling on" | | 13 | "The victim lay in a" | | 14 | "The tunnel floor beyond the" | | 15 | "Quinn asked the nearest forensics" | | 16 | "The young man looked uncomfortable" | | 17 | "Quinn finished for him" | | 18 | "She'd already noticed." | | 19 | "Every entrance to this section" |
| | ratio | 0.79 | |
| 100.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 2 | | totalSentences | 81 | | matches | | 0 | "Before Eva could answer, one" | | 1 | "To whatever waited in that" |
| | ratio | 0.025 | |
| 98.21% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 32 | | technicalSentenceCount | 2 | | matches | | 0 | "The victim lay in a perfect circle of undisturbed dust, as if he'd been gently placed there by invisible hands." | | 1 | "To whatever waited in that darkness that swallowed light and left bodies in impossible circles." |
| |
| 34.09% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 22 | | uselessAdditionCount | 4 | | matches | | 0 | "Eva Kowalski called, her voice echoing off the curved walls" | | 1 | "Quinn stood, her brown eyes scanning the walls" | | 2 | "She paused, seeming to weigh her words" | | 3 | "Quinn's voice had, professional" |
| |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 8 | | fancyCount | 1 | | fancyTags | | 0 | "the officer confirmed (confirm)" |
| | dialogueSentences | 43 | | tagDensity | 0.186 | | leniency | 0.372 | | rawRatio | 0.125 | | effectiveRatio | 0.047 | |