| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 1 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 32 | | tagDensity | 0.031 | | leniency | 0.063 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 97.62% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 2098 | | 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) | |
| 88.08% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 2098 | | totalAiIsms | 5 | | found | | | highlights | | 0 | "lurched" | | 1 | "pulse" | | 2 | "velvet" | | 3 | "flicker" | | 4 | "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 | 215 | | matches | | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 1 | | hedgeCount | 1 | | narrationSentences | 215 | | filterMatches | | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 246 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 37 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 2098 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 0 | | unquotedAttributions | 0 | | matches | (empty) | |
| 50.00% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 81 | | wordCount | 1908 | | uniqueNames | 25 | | maxNameDensity | 1.47 | | worstName | "Quinn" | | maxWindowNameDensity | 3.5 | | worstWindowName | "Quinn" | | discoveredNames | | Soho | 1 | | Raven | 1 | | Nest | 1 | | Harlow | 1 | | Quinn | 28 | | London | 1 | | Vale | 14 | | Greek | 1 | | Street | 1 | | Charing | 1 | | Cross | 1 | | Road | 1 | | Deptford | 1 | | Morris | 2 | | TfL | 1 | | Veil | 1 | | Market | 1 | | Saint | 1 | | Christopher | 1 | | Herrera | 1 | | Tomás | 7 | | Rain | 4 | | Two | 3 | | One | 3 | | Metal | 3 |
| | persons | | 0 | "Nest" | | 1 | "Harlow" | | 2 | "Quinn" | | 3 | "Vale" | | 4 | "Morris" | | 5 | "Saint" | | 6 | "Christopher" | | 7 | "Herrera" | | 8 | "Tomás" | | 9 | "Rain" | | 10 | "Metal" |
| | places | | 0 | "Soho" | | 1 | "Raven" | | 2 | "London" | | 3 | "Greek" | | 4 | "Street" | | 5 | "Charing" | | 6 | "Cross" | | 7 | "Road" | | 8 | "Deptford" |
| | globalScore | 0.766 | | windowScore | 0.5 | |
| 93.18% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 132 | | glossingSentenceCount | 3 | | matches | | 0 | "as if testing it" | | 1 | "looked like the sort of woman people lied" | | 2 | "not quite amusement" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 1 | | per1kWords | 0.477 | | wordCount | 2098 | | matches | | 0 | "not properly, but the place announced itself with the confidence of something" |
| |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 1 | | totalSentences | 246 | | matches | | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 105 | | mean | 19.98 | | std | 23.05 | | cv | 1.154 | | sampleLengths | | 0 | 80 | | 1 | 102 | | 2 | 64 | | 3 | 3 | | 4 | 44 | | 5 | 5 | | 6 | 52 | | 7 | 5 | | 8 | 4 | | 9 | 8 | | 10 | 1 | | 11 | 5 | | 12 | 54 | | 13 | 37 | | 14 | 1 | | 15 | 10 | | 16 | 55 | | 17 | 29 | | 18 | 3 | | 19 | 60 | | 20 | 22 | | 21 | 2 | | 22 | 5 | | 23 | 2 | | 24 | 49 | | 25 | 39 | | 26 | 64 | | 27 | 5 | | 28 | 5 | | 29 | 42 | | 30 | 11 | | 31 | 41 | | 32 | 5 | | 33 | 48 | | 34 | 24 | | 35 | 34 | | 36 | 7 | | 37 | 7 | | 38 | 2 | | 39 | 14 | | 40 | 78 | | 41 | 9 | | 42 | 21 | | 43 | 102 | | 44 | 3 | | 45 | 79 | | 46 | 6 | | 47 | 4 | | 48 | 40 | | 49 | 27 |
| |
| 100.00% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 2 | | totalSentences | 215 | | matches | | 0 | "been closed" | | 1 | "been covered" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 300 | | matches | (empty) | |
| 100.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 0 | | semicolonCount | 0 | | flaggedSentences | 0 | | totalSentences | 246 | | ratio | 0 | | matches | (empty) | |
| 94.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 1917 | | adjectiveStacks | 1 | | stackExamples | | 0 | "pressing hard against her" |
| | adverbCount | 51 | | adverbRatio | 0.026604068857589983 | | lyAdverbCount | 9 | | lyAdverbRatio | 0.004694835680751174 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 246 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 246 | | mean | 8.53 | | std | 6.77 | | cv | 0.794 | | sampleLengths | | 0 | 11 | | 1 | 19 | | 2 | 31 | | 3 | 7 | | 4 | 7 | | 5 | 5 | | 6 | 4 | | 7 | 7 | | 8 | 12 | | 9 | 7 | | 10 | 10 | | 11 | 12 | | 12 | 7 | | 13 | 6 | | 14 | 6 | | 15 | 6 | | 16 | 2 | | 17 | 3 | | 18 | 20 | | 19 | 8 | | 20 | 1 | | 21 | 8 | | 22 | 2 | | 23 | 2 | | 24 | 7 | | 25 | 9 | | 26 | 21 | | 27 | 6 | | 28 | 3 | | 29 | 8 | | 30 | 2 | | 31 | 2 | | 32 | 13 | | 33 | 8 | | 34 | 11 | | 35 | 5 | | 36 | 22 | | 37 | 6 | | 38 | 24 | | 39 | 5 | | 40 | 4 | | 41 | 4 | | 42 | 2 | | 43 | 2 | | 44 | 1 | | 45 | 5 | | 46 | 14 | | 47 | 4 | | 48 | 13 | | 49 | 9 |
| |
| 69.65% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 7 | | diversityRatio | 0.4349593495934959 | | totalSentences | 246 | | uniqueOpeners | 107 | |
| 100.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 7 | | totalSentences | 187 | | matches | | 0 | "Only darkness descending under a" | | 1 | "Somewhere nearby, a bottle smashed." | | 2 | "Then she pictured Vale slipping" | | 3 | "Instead it blazed." | | 4 | "Then he disappeared down the" | | 5 | "Then came a sound Quinn" | | 6 | "Then she looked at the" |
| | ratio | 0.037 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 30 | | totalSentences | 187 | | matches | | 0 | "Her brown eyes stayed on" | | 1 | "Her leather watch strap had" | | 2 | "She had watched the place" | | 3 | "They knocked once, walked in," | | 4 | "He paused with his face" | | 5 | "He started east without an" | | 6 | "She crossed through traffic between" | | 7 | "Her shoes slapped through thin" | | 8 | "He looked back once." | | 9 | "He broke into a run." | | 10 | "Her breath came steady." | | 11 | "He moved like he knew" | | 12 | "She had seen that before." | | 13 | "She had found Morris an" | | 14 | "She looked like the sort" | | 15 | "He reached the mouth of" | | 16 | "She drew a breath and" | | 17 | "She stared into the stairwell." | | 18 | "She jammed the phone away," | | 19 | "She descended the last flight" |
| | ratio | 0.16 | |
| 77.65% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 143 | | totalSentences | 187 | | matches | | 0 | "Rain needled across Soho and" | | 1 | "Detective Harlow Quinn stood under" | | 2 | "Her brown eyes stayed on" | | 3 | "Her leather watch strap had" | | 4 | "The second hand kept moving." | | 5 | "Smokers had crushed cigarettes into" | | 6 | "The green sign hummed above" | | 7 | "She had watched the place" | | 8 | "Students came in loud and" | | 9 | "Office workers drifted through at" | | 10 | "Men and women on their" | | 11 | "They knocked once, walked in," | | 12 | "Quinn checked the photo on" | | 13 | "A man in profile beneath" | | 14 | "A canvas satchel slung across" | | 15 | "Elias Vale, known broker, suspected" | | 16 | "The last lead had pointed" | | 17 | "The door opened." | | 18 | "A man stepped out beneath" | | 19 | "He paused with his face" |
| | ratio | 0.765 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 187 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 77 | | technicalSentenceCount | 3 | | matches | | 0 | "Linked to three violent robberies, one missing witness, and a morgue attendant who had burned to death in a locked room." | | 1 | "He stood behind a table laid out with bandages, ampoules, syringes, brown bottles, a blood pressure cuff that had seen better decades." | | 2 | "Then she looked at the tunnel mouth where Vale had vanished, at the market in full swing behind her, at the old station remade into something that had grown ben…" |
| |
| 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 | |