| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 8 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 29 | | tagDensity | 0.276 | | leniency | 0.552 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 95.51% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1114 | | 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) | |
| 37.16% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1114 | | totalAiIsms | 14 | | found | | 0 | | | 1 | | | 2 | | | 3 | | | 4 | | | 5 | | | 6 | | | 7 | | | 8 | | | 9 | | | 10 | | | 11 | | word | "down her spine" | | count | 1 |
|
| | highlights | | 0 | "flicker" | | 1 | "pulse" | | 2 | "stomach" | | 3 | "unreadable" | | 4 | "weight" | | 5 | "echoed" | | 6 | "velvet" | | 7 | "scanned" | | 8 | "familiar" | | 9 | "tracing" | | 10 | "raced" | | 11 | "down her spine" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 1 | | maxInWindow | 1 | | found | | 0 | | label | "stomach dropped/sank" | | count | 1 |
|
| | highlights | | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 0 | | narrationSentences | 125 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 3 | | hedgeCount | 0 | | narrationSentences | 125 | | filterMatches | | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 146 | | 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 | 1094 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 7 | | unquotedAttributions | 0 | | matches | (empty) | |
| 83.33% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 19 | | wordCount | 854 | | uniqueNames | 11 | | maxNameDensity | 0.59 | | worstName | "Tomás" | | maxWindowNameDensity | 2.5 | | worstWindowName | "Tomás" | | discoveredNames | | Soho | 1 | | Harlow | 1 | | Quinn | 4 | | Morris | 2 | | Veil | 1 | | Market | 1 | | Camden | 1 | | Underground | 1 | | Saint | 1 | | Christopher | 1 | | Tomás | 5 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Morris" | | 3 | "Market" | | 4 | "Saint" | | 5 | "Christopher" | | 6 | "Tomás" |
| | places | | | globalScore | 1 | | windowScore | 0.833 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 55 | | glossingSentenceCount | 0 | | matches | (empty) | |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1094 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 146 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 65 | | mean | 16.83 | | std | 14.64 | | cv | 0.87 | | sampleLengths | | 0 | 61 | | 1 | 48 | | 2 | 24 | | 3 | 9 | | 4 | 4 | | 5 | 57 | | 6 | 8 | | 7 | 18 | | 8 | 35 | | 9 | 9 | | 10 | 7 | | 11 | 51 | | 12 | 15 | | 13 | 2 | | 14 | 7 | | 15 | 27 | | 16 | 4 | | 17 | 57 | | 18 | 39 | | 19 | 15 | | 20 | 12 | | 21 | 35 | | 22 | 2 | | 23 | 12 | | 24 | 2 | | 25 | 31 | | 26 | 14 | | 27 | 6 | | 28 | 3 | | 29 | 29 | | 30 | 7 | | 31 | 5 | | 32 | 6 | | 33 | 10 | | 34 | 15 | | 35 | 7 | | 36 | 15 | | 37 | 6 | | 38 | 3 | | 39 | 7 | | 40 | 37 | | 41 | 3 | | 42 | 5 | | 43 | 24 | | 44 | 14 | | 45 | 5 | | 46 | 18 | | 47 | 15 | | 48 | 5 | | 49 | 20 |
| |
| 100.00% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 0 | | totalSentences | 125 | | matches | (empty) | |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 1 | | totalVerbs | 156 | | matches | | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 12 | | semicolonCount | 0 | | flaggedSentences | 9 | | totalSentences | 146 | | ratio | 0.062 | | matches | | 0 | "She rounded the corner, coat flaring behind her like the wings of a wounded bird, eyes locked on the figure ahead—half-shadow, half-silhouette—darting between overflowing dumpsters and sagging fire escapes." | | 1 | "He wasn’t just running from her—he was running from something else too." | | 2 | "She hit the dead end—brick wall, graffiti-tagged bin, fire door chained shut—and for a second, her stomach dropped." | | 3 | "She remembered the file Morris had hidden—notes scrawled in the margins about “the Market,” about movement under the full moon." | | 4 | "Her watch ticked—too loud." | | 5 | "A stall to her left sold teeth—human, animal, things with too many roots." | | 6 | "Then—a flicker." | | 7 | "Her watch—her father’s watch—shivered on her wrist." | | 8 | "Somewhere in the tunnels, a bell rang—deep, hollow, final." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 872 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 25 | | adverbRatio | 0.028669724770642203 | | lyAdverbCount | 3 | | lyAdverbRatio | 0.0034403669724770644 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 146 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 146 | | mean | 7.49 | | std | 6.22 | | cv | 0.831 | | sampleLengths | | 0 | 17 | | 1 | 15 | | 2 | 29 | | 3 | 5 | | 4 | 2 | | 5 | 24 | | 6 | 12 | | 7 | 5 | | 8 | 12 | | 9 | 2 | | 10 | 4 | | 11 | 3 | | 12 | 3 | | 13 | 9 | | 14 | 4 | | 15 | 20 | | 16 | 6 | | 17 | 12 | | 18 | 8 | | 19 | 3 | | 20 | 2 | | 21 | 2 | | 22 | 4 | | 23 | 8 | | 24 | 18 | | 25 | 6 | | 26 | 1 | | 27 | 7 | | 28 | 21 | | 29 | 9 | | 30 | 7 | | 31 | 3 | | 32 | 20 | | 33 | 6 | | 34 | 1 | | 35 | 5 | | 36 | 16 | | 37 | 11 | | 38 | 4 | | 39 | 2 | | 40 | 7 | | 41 | 12 | | 42 | 3 | | 43 | 5 | | 44 | 2 | | 45 | 1 | | 46 | 4 | | 47 | 4 | | 48 | 14 | | 49 | 16 |
| |
| 50.91% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 14 | | diversityRatio | 0.363013698630137 | | totalSentences | 146 | | uniqueOpeners | 53 | |
| 100.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 8 | | totalSentences | 98 | | matches | | 0 | "Too many witnesses." | | 1 | "Too many windows." | | 2 | "Then she saw it: the" | | 3 | "Then the corridor opened." | | 4 | "Then someone stepped from the" | | 5 | "Then stepped closer." | | 6 | "Somewhere in the tunnels, a" | | 7 | "Then at the last flicker" |
| | ratio | 0.082 | |
| 85.31% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 33 | | totalSentences | 98 | | matches | | 0 | "She rounded the corner, coat" | | 1 | "She didn’t know his name." | | 2 | "He wasn’t just running from" | | 3 | "Her fingers brushed the butt" | | 4 | "she shouted, voice sharp over" | | 5 | "He didn’t look back." | | 6 | "Her pulse hammered behind her" | | 7 | "She hit the dead end—brick" | | 8 | "She dropped to one knee," | | 9 | "She remembered the file Morris" | | 10 | "She’d thought it was grief" | | 11 | "She’d found it wrapped in" | | 12 | "She reached into her coat," | | 13 | "She pocketed the token and" | | 14 | "Her boots echoed like gunshots." | | 15 | "Her watch ticked—too loud." | | 16 | "She scanned the crowd." | | 17 | "She passed a vendor selling" | | 18 | "She stopped, breathing hard." | | 19 | "Her watch—her father’s watch—shivered on" |
| | ratio | 0.337 | |
| 100.00% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 70 | | totalSentences | 98 | | matches | | 0 | "Rain slapped the pavement like" | | 1 | "Detective Harlow Quinn’s boots skidded" | | 2 | "She rounded the corner, coat" | | 3 | "She didn’t know his name." | | 4 | "He wasn’t just running from" | | 5 | "Something deeper than the law." | | 6 | "Her fingers brushed the butt" | | 7 | "she shouted, voice sharp over" | | 8 | "He didn’t look back." | | 9 | "Quinn tightened her grip on" | | 10 | "Her pulse hammered behind her" | | 11 | "This felt wrong." | | 12 | "Every instinct screamed trap." | | 13 | "She hit the dead end—brick" | | 14 | "A breath of warmer air" | | 15 | "She dropped to one knee," | | 16 | "The bone token." | | 17 | "She remembered the file Morris" | | 18 | "She’d thought it was grief" | | 19 | "She’d found it wrapped in" |
| | ratio | 0.714 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 98 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 32 | | technicalSentenceCount | 0 | | matches | (empty) | |
| 62.50% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 8 | | uselessAdditionCount | 1 | | matches | | 0 | "she shouted, voice sharp over the storm" |
| |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 6 | | fancyCount | 1 | | fancyTags | | | dialogueSentences | 29 | | tagDensity | 0.207 | | leniency | 0.414 | | rawRatio | 0.167 | | effectiveRatio | 0.069 | |