| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 1 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 2 | | tagDensity | 0.5 | | leniency | 1 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 95.74% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1174 | | totalAiIsmAdverbs | 1 | | found | | | highlights | | |
| 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) | |
| 23.34% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1174 | | totalAiIsms | 18 | | found | | 0 | | | 1 | | | 2 | | | 3 | | | 4 | | | 5 | | | 6 | | | 7 | | | 8 | | | 9 | | | 10 | | | 11 | | | 12 | | | 13 | | | 14 | | | 15 | | word | "down her spine" | | count | 1 |
|
| | highlights | | 0 | "scanned" | | 1 | "glinting" | | 2 | "flicker" | | 3 | "pulse" | | 4 | "calculating" | | 5 | "echoing" | | 6 | "weight" | | 7 | "dancing" | | 8 | "pulsed" | | 9 | "stomach" | | 10 | "etched" | | 11 | "quickened" | | 12 | "silence" | | 13 | "flickered" | | 14 | "chill" | | 15 | "down her spine" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 1 | | maxInWindow | 1 | | found | | 0 | | label | "sent a shiver through" | | count | 1 |
|
| | highlights | | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 1 | | narrationSentences | 113 | | matches | | |
| 92.29% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 1 | | hedgeCount | 3 | | narrationSentences | 113 | | filterMatches | | | hedgeMatches | | 0 | "seemed to" | | 1 | "happened to" |
| |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 114 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 32 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 9 | | markdownWords | 19 | | totalWords | 1160 | | ratio | 0.016 | | matches | | 0 | "The Raven’s Nest" | | 1 | "wrong" | | 2 | "actual torches" | | 3 | "teeth" | | 4 | "couldn’t" | | 5 | "eyeballs" | | 6 | "was" | | 7 | "\"Quinn, something’s wrong with this one.\"" | | 8 | "\"You’re late, Tomás.\"" |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 1 | | unquotedAttributions | 0 | | matches | (empty) | |
| 97.87% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 32 | | wordCount | 1151 | | uniqueNames | 12 | | maxNameDensity | 1.04 | | worstName | "Quinn" | | maxWindowNameDensity | 1.5 | | worstWindowName | "Quinn" | | discoveredNames | | London | 1 | | Harlow | 1 | | Quinn | 12 | | Soho | 1 | | Raven | 1 | | Tomás | 1 | | Herrera | 9 | | Saint | 1 | | Christopher | 1 | | Camden | 1 | | Tube | 1 | | Morris | 2 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Tomás" | | 3 | "Herrera" | | 4 | "Saint" | | 5 | "Christopher" | | 6 | "Camden" | | 7 | "Morris" |
| | places | | | globalScore | 0.979 | | windowScore | 1 | |
| 47.26% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 73 | | glossingSentenceCount | 3 | | matches | | 0 | "looked like *teeth*—but not human" | | 1 | "looked like *eyeballs* floating in brine" | | 2 | "shadows that seemed to move on their own" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1160 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 1 | | totalSentences | 114 | | matches | | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 30 | | mean | 38.67 | | std | 27.03 | | cv | 0.699 | | sampleLengths | | 0 | 80 | | 1 | 34 | | 2 | 78 | | 3 | 29 | | 4 | 74 | | 5 | 4 | | 6 | 53 | | 7 | 32 | | 8 | 53 | | 9 | 41 | | 10 | 54 | | 11 | 56 | | 12 | 104 | | 13 | 29 | | 14 | 30 | | 15 | 71 | | 16 | 54 | | 17 | 9 | | 18 | 33 | | 19 | 42 | | 20 | 3 | | 21 | 57 | | 22 | 4 | | 23 | 66 | | 24 | 8 | | 25 | 31 | | 26 | 3 | | 27 | 14 | | 28 | 10 | | 29 | 4 |
| |
| 99.05% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 2 | | totalSentences | 113 | | matches | | 0 | "was rusted" | | 1 | "was carved" |
| |
| 61.83% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 4 | | totalVerbs | 193 | | matches | | 0 | "was winning" | | 1 | "was closing" | | 2 | "was pressing" | | 3 | "wasn’t running" |
| |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 14 | | semicolonCount | 0 | | flaggedSentences | 11 | | totalSentences | 114 | | ratio | 0.096 | | matches | | 0 | "She’d been tailing the suspect for blocks—ever since he’d bolted from *The Raven’s Nest* like a man with something to hide." | | 1 | "The suspect—Tomás Herrera, if her sources were right—was fast, but she was faster." | | 2 | "She knew these streets—had walked them a hundred times in the day, a hundred more at night." | | 3 | "She knew Camden’s underbelly—knew the abandoned Tube stations, the forgotten tunnels, the places where the city’s secrets festered." | | 4 | "Flickering torches—*actual torches*—cast long, dancing shadows along the curved ceiling." | | 5 | "A man in a long coat haggled over a vial of something black and viscous, while a woman with too many teeth examined a necklace of what looked like *teeth*—but not human." | | 6 | "A child—if it *was* a child—with too-long limbs and a grin full of needle teeth giggled as she passed." | | 7 | "The vendor, a gaunt figure with fingers like spider legs, handed him a single token—a small, yellowed bone etched with strange symbols." | | 8 | "Follow him deeper into this nightmare, or retreat and call for backup—knowing full well that by the time anyone arrived, the market would be gone." | | 9 | "Not a modern one—this was ancient, its wood blackened with age, its iron hinges rusted." | | 10 | "Another voice answered—deeper, older." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 1165 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 22 | | adverbRatio | 0.01888412017167382 | | lyAdverbCount | 5 | | lyAdverbRatio | 0.004291845493562232 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 114 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 114 | | mean | 10.18 | | std | 6.82 | | cv | 0.67 | | sampleLengths | | 0 | 15 | | 1 | 23 | | 2 | 18 | | 3 | 24 | | 4 | 21 | | 5 | 13 | | 6 | 13 | | 7 | 10 | | 8 | 6 | | 9 | 30 | | 10 | 11 | | 11 | 5 | | 12 | 3 | | 13 | 16 | | 14 | 13 | | 15 | 14 | | 16 | 8 | | 17 | 17 | | 18 | 4 | | 19 | 16 | | 20 | 15 | | 21 | 4 | | 22 | 15 | | 23 | 9 | | 24 | 3 | | 25 | 19 | | 26 | 7 | | 27 | 5 | | 28 | 3 | | 29 | 18 | | 30 | 2 | | 31 | 4 | | 32 | 13 | | 33 | 12 | | 34 | 10 | | 35 | 2 | | 36 | 2 | | 37 | 14 | | 38 | 3 | | 39 | 5 | | 40 | 3 | | 41 | 7 | | 42 | 11 | | 43 | 6 | | 44 | 6 | | 45 | 13 | | 46 | 9 | | 47 | 5 | | 48 | 14 | | 49 | 13 |
| |
| 35.96% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 16 | | diversityRatio | 0.24561403508771928 | | totalSentences | 114 | | uniqueOpeners | 28 | |
| 100.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 5 | | totalSentences | 104 | | matches | | 0 | "Then she saw it." | | 1 | "Just the steady drip of" | | 2 | "Just like the last time." | | 3 | "Just like Morris." | | 4 | "Then she saw it." |
| | ratio | 0.048 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 28 | | totalSentences | 104 | | matches | | 0 | "Her breath came in controlled" | | 1 | "She’d been tailing the suspect" | | 2 | "He had youth on his" | | 3 | "She could see him ahead," | | 4 | "His olive skin was pale" | | 5 | "He wasn’t built for running." | | 6 | "She didn’t need to check" | | 7 | "She knew these streets—had walked" | | 8 | "He kicked the grate open" | | 9 | "He dropped inside without a" | | 10 | "She knew Camden’s underbelly—knew the" | | 11 | "She crouched at the edge" | | 12 | "She could call it in." | | 13 | "She swung her legs over" | | 14 | "She moved forward, her steps" | | 15 | "She’d seen things in her" | | 16 | "She spotted Herrera ahead, weaving" | | 17 | "He wasn’t running now." | | 18 | "He moved with purpose, nodding" | | 19 | "He was known here." |
| | ratio | 0.269 | |
| 60.96% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 83 | | totalSentences | 104 | | matches | | 0 | "The rain came down in" | | 1 | "Detective Harlow Quinn moved with" | | 2 | "The neon glow of Soho’s" | | 3 | "Her breath came in controlled" | | 4 | "She’d been tailing the suspect" | | 5 | "The suspect—Tomás Herrera, if her" | | 6 | "He had youth on his" | | 7 | "She could see him ahead," | | 8 | "His olive skin was pale" | | 9 | "He wasn’t built for running." | | 10 | "Quinn adjusted the worn leather" | | 11 | "She didn’t need to check" | | 12 | "Herrera ducked into a narrow" | | 13 | "Quinn followed, her pulse steady," | | 14 | "She knew these streets—had walked" | | 15 | "The air smelled different here," | | 16 | "The rain seemed to fall" | | 17 | "A grate in the pavement," | | 18 | "The metal was rusted, but" | | 19 | "Herrera didn’t hesitate." |
| | ratio | 0.798 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 104 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 53 | | technicalSentenceCount | 2 | | matches | | 0 | "Detective Harlow Quinn moved with the precision of a soldier, her boots splashing through puddles that pooled in the cracks of the pavement." | | 1 | "The torchlight flickered, casting monstrous shadows that seemed to move on their own." |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 1 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | |