| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 12 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 33 | | tagDensity | 0.364 | | leniency | 0.727 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 71.54% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1054 | | totalAiIsmAdverbs | 6 | | found | | 0 | | | 1 | | | 2 | | | 3 | | | 4 | | adverb | "deliberately" | | count | 1 |
|
| | highlights | | 0 | "precisely" | | 1 | "carefully" | | 2 | "perfectly" | | 3 | "very" | | 4 | "deliberately" |
| |
| 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) | |
| 76.28% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1054 | | totalAiIsms | 5 | | found | | | highlights | | 0 | "marble" | | 1 | "navigated" | | 2 | "measured" | | 3 | "silence" |
| |
| 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 | 0 | | narrationSentences | 62 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 0 | | hedgeCount | 1 | | narrationSentences | 62 | | filterMatches | (empty) | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 83 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 64 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 1 | | markdownWords | 14 | | totalWords | 1047 | | ratio | 0.013 | | matches | | 0 | "fine, go, but I will be listening from the kitchen with the door open." |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 21 | | unquotedAttributions | 0 | | matches | (empty) | |
| 83.33% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 36 | | wordCount | 817 | | uniqueNames | 10 | | maxNameDensity | 1.22 | | worstName | "Rory" | | maxWindowNameDensity | 2.5 | | worstWindowName | "Eva" | | discoveredNames | | Ptolemy | 3 | | Rory | 10 | | Moreau | 2 | | Eva | 10 | | Bermondsey | 1 | | Lucien | 6 | | Brick | 1 | | Lane | 1 | | London | 1 | | One | 1 |
| | persons | | 0 | "Ptolemy" | | 1 | "Rory" | | 2 | "Moreau" | | 3 | "Eva" | | 4 | "Lucien" |
| | places | | 0 | "Brick" | | 1 | "Lane" | | 2 | "London" | | 3 | "One" |
| | globalScore | 0.888 | | windowScore | 0.833 | |
| 0.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 37 | | glossingSentenceCount | 5 | | matches | | 0 | "something between dread and relief told Rory ev" | | 1 | "der, which Eva apparently read correctly as *" | | 2 | "quite discomfort but was the closest she'd seen him get to it" | | 3 | "Not quite a smile, but the architecture of one" | | 4 | "not quite touching, the gap between them precisely and deliberately maintained" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 1 | | per1kWords | 0.955 | | wordCount | 1047 | | matches | | 0 | "Not quite a smile, but the architecture of one" |
| |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 83 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 46 | | mean | 22.76 | | std | 22.21 | | cv | 0.976 | | sampleLengths | | 0 | 23 | | 1 | 64 | | 2 | 68 | | 3 | 12 | | 4 | 6 | | 5 | 12 | | 6 | 6 | | 7 | 81 | | 8 | 81 | | 9 | 12 | | 10 | 4 | | 11 | 34 | | 12 | 44 | | 13 | 37 | | 14 | 5 | | 15 | 6 | | 16 | 4 | | 17 | 10 | | 18 | 6 | | 19 | 13 | | 20 | 37 | | 21 | 11 | | 22 | 23 | | 23 | 29 | | 24 | 9 | | 25 | 43 | | 26 | 3 | | 27 | 4 | | 28 | 52 | | 29 | 2 | | 30 | 3 | | 31 | 29 | | 32 | 25 | | 33 | 66 | | 34 | 4 | | 35 | 2 | | 36 | 19 | | 37 | 20 | | 38 | 5 | | 39 | 13 | | 40 | 18 | | 41 | 4 | | 42 | 41 | | 43 | 49 | | 44 | 3 | | 45 | 5 |
| |
| 93.94% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 2 | | totalSentences | 62 | | matches | | 0 | "was, perched" | | 1 | "being filled" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 141 | | matches | (empty) | |
| 100.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 0 | | semicolonCount | 0 | | flaggedSentences | 0 | | totalSentences | 83 | | ratio | 0 | | matches | (empty) | |
| 91.79% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 823 | | adjectiveStacks | 1 | | stackExamples | | 0 | "strange two-toned quality" |
| | adverbCount | 35 | | adverbRatio | 0.0425273390036452 | | lyAdverbCount | 16 | | lyAdverbRatio | 0.019441069258809233 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 83 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 83 | | mean | 12.61 | | std | 11.81 | | cv | 0.937 | | sampleLengths | | 0 | 23 | | 1 | 34 | | 2 | 5 | | 3 | 25 | | 4 | 12 | | 5 | 11 | | 6 | 24 | | 7 | 21 | | 8 | 12 | | 9 | 6 | | 10 | 8 | | 11 | 4 | | 12 | 3 | | 13 | 3 | | 14 | 45 | | 15 | 36 | | 16 | 20 | | 17 | 7 | | 18 | 41 | | 19 | 3 | | 20 | 10 | | 21 | 12 | | 22 | 4 | | 23 | 26 | | 24 | 8 | | 25 | 4 | | 26 | 40 | | 27 | 2 | | 28 | 17 | | 29 | 11 | | 30 | 7 | | 31 | 5 | | 32 | 6 | | 33 | 4 | | 34 | 5 | | 35 | 5 | | 36 | 6 | | 37 | 13 | | 38 | 13 | | 39 | 19 | | 40 | 5 | | 41 | 11 | | 42 | 2 | | 43 | 21 | | 44 | 29 | | 45 | 4 | | 46 | 5 | | 47 | 32 | | 48 | 11 | | 49 | 3 |
| |
| 61.04% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 6 | | diversityRatio | 0.40963855421686746 | | totalSentences | 83 | | uniqueOpeners | 34 | |
| 0.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 0 | | totalSentences | 51 | | matches | (empty) | | ratio | 0 | |
| 47.45% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 22 | | totalSentences | 51 | | matches | | 0 | "She clearly expected a delivery." | | 1 | "His mismatched eyes settled on" | | 2 | "He didn't move from the" | | 3 | "She hadn't seen him in" | | 4 | "He'd said nothing." | | 5 | "His amber eye caught hers" | | 6 | "He sat at the man's" | | 7 | "He turned the cane once" | | 8 | "he corrected, and the precision" | | 9 | "She hadn't expected that." | | 10 | "She kept her face still." | | 11 | "His gaze held hers, steady," | | 12 | "She needed the height advantage," | | 13 | "She crossed to the window," | | 14 | "His jaw had shifted, something" | | 15 | "he said, simply, and somehow" | | 16 | "She crossed her arms." | | 17 | "He reached into the inner" | | 18 | "She crossed the room and" | | 19 | "She'd seen it before." |
| | ratio | 0.431 | |
| 9.02% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 46 | | totalSentences | 51 | | matches | | 0 | "The knock came at half" | | 1 | "Eva yanked open the door" | | 2 | "She clearly expected a delivery." | | 3 | "The way her expression collapsed" | | 4 | "Lucien Moreau filled the doorway" | | 5 | "Charcoal suit, not a crease" | | 6 | "The ivory-handled cane rested at" | | 7 | "His mismatched eyes settled on" | | 8 | "The word landed between them" | | 9 | "He didn't move from the" | | 10 | "Eva stepped aside." | | 11 | "Ptolemy did not." | | 12 | "The flat closed around him" | | 13 | "Lucien navigated it without touching" | | 14 | "Rory stayed where she was," | | 15 | "She hadn't seen him in" | | 16 | "The last time, they'd stood" | | 17 | "He'd said nothing." | | 18 | "That had been worse than" | | 19 | "His amber eye caught hers" |
| | ratio | 0.902 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 51 | | matches | (empty) | | ratio | 0 | |
| 60.44% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 26 | | technicalSentenceCount | 3 | | matches | | 0 | "Lucien navigated it without touching a single thing, which was its own kind of skill, and settled into the armchair by the window as though he'd measured it for…" | | 1 | "His jaw had shifted, something settling there that wasn't quite discomfort but was the closest she'd seen him get to it." | | 2 | "Lucien Moreau, who spoke four languages and moved through London's supernatural underworld like oil across water, who fixed unfixable problems for people who di…" |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 12 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 8 | | fancyCount | 1 | | fancyTags | | 0 | "he corrected (correct)" |
| | dialogueSentences | 33 | | tagDensity | 0.242 | | leniency | 0.485 | | rawRatio | 0.125 | | effectiveRatio | 0.061 | |