| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 2 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 18 | | tagDensity | 0.111 | | leniency | 0.222 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 100.00% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 452 | | totalAiIsmAdverbs | 0 | | found | (empty) | | highlights | (empty) | |
| 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) | |
| 33.63% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 452 | | totalAiIsms | 6 | | found | | | highlights | | 0 | "flickered" | | 1 | "weight" | | 2 | "measured" | | 3 | "glinting" | | 4 | "implication" | | 5 | "flicked" |
| |
| 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 | 38 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 0 | | hedgeCount | 0 | | narrationSentences | 38 | | filterMatches | (empty) | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 54 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 30 | | ratio | 0 | | matches | (empty) | |
| 0.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 18 | | markdownWords | 100 | | totalWords | 445 | | ratio | 0.225 | | matches | | 0 | "\"You’re late.\"" | | 1 | "\"You’re the one who left the trail thin.\"" | | 2 | "\"I didn’t leave a trail. You followed it.\"" | | 3 | "\"You’re selling something at the Market.\"" | | 4 | "\"Yes.\"" | | 5 | "\"And you’re not alone.\"" | | 6 | "\"You’re not supposed to be here.\"" | | 7 | "\"I’m not supposed to be anywhere near what you’re dealing.\"" | | 8 | "\"Then why are you?\"" | | 9 | "\"Because I’m a cop. And because I know you’re not just selling potions.\"" | | 10 | "\"And what makes you think that?\"" | | 11 | "\"Because you’re not selling them in the Nest.\"" | | 12 | "\"Because they’re not for the Nest.\"" | | 13 | "\"You’re not supposed to be here,\"" | | 14 | "\"But you’re here now.\"" | | 15 | "\"You’ll be careful,\"" | | 16 | "\"I’ll be smart.\"" | | 17 | "\"Then don’t.\"" |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 3 | | unquotedAttributions | 0 | | matches | (empty) | |
| 34.06% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 19 | | wordCount | 345 | | uniqueNames | 8 | | maxNameDensity | 2.32 | | worstName | "Quinn" | | maxWindowNameDensity | 3 | | worstWindowName | "Quinn" | | discoveredNames | | Raven | 1 | | Nest | 1 | | Harlow | 1 | | Quinn | 8 | | Soho | 1 | | Tomás | 4 | | Herrera | 1 | | Market | 2 |
| | persons | | 0 | "Raven" | | 1 | "Nest" | | 2 | "Harlow" | | 3 | "Quinn" | | 4 | "Tomás" | | 5 | "Herrera" | | 6 | "Market" |
| | places | | | globalScore | 0.341 | | windowScore | 0.667 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 23 | | 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 | 445 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 54 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 26 | | mean | 17.12 | | std | 20.58 | | cv | 1.202 | | sampleLengths | | 0 | 79 | | 1 | 56 | | 2 | 35 | | 3 | 2 | | 4 | 13 | | 5 | 17 | | 6 | 6 | | 7 | 1 | | 8 | 4 | | 9 | 17 | | 10 | 6 | | 11 | 10 | | 12 | 4 | | 13 | 19 | | 14 | 6 | | 15 | 8 | | 16 | 6 | | 17 | 67 | | 18 | 15 | | 19 | 7 | | 20 | 40 | | 21 | 5 | | 22 | 3 | | 23 | 2 | | 24 | 12 | | 25 | 5 |
| |
| 100.00% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 0 | | totalSentences | 38 | | matches | (empty) | |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 54 | | matches | (empty) | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 6 | | semicolonCount | 0 | | flaggedSentences | 4 | | totalSentences | 54 | | ratio | 0.074 | | matches | | 0 | "The suspect had vanished into the labyrinth of Soho’s back alleys, leaving only the scent of damp pavement and something sharper—burnt oil, maybe, or the metallic tang of fear." | | 1 | "The man—Tomás Herrera—kicked a loose stone into the gutter." | | 2 | "But the suspect—someone she’d been tracking for days—had led her here." | | 3 | "The air below was colder, thicker, the scent of damp stone and something older—something that made her skin prickle." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 352 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 10 | | adverbRatio | 0.028409090909090908 | | lyAdverbCount | 1 | | lyAdverbRatio | 0.002840909090909091 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 54 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 54 | | mean | 8.24 | | std | 5.61 | | cv | 0.681 | | sampleLengths | | 0 | 17 | | 1 | 22 | | 2 | 11 | | 3 | 29 | | 4 | 9 | | 5 | 20 | | 6 | 7 | | 7 | 7 | | 8 | 13 | | 9 | 10 | | 10 | 14 | | 11 | 11 | | 12 | 2 | | 13 | 5 | | 14 | 8 | | 15 | 9 | | 16 | 8 | | 17 | 6 | | 18 | 1 | | 19 | 4 | | 20 | 2 | | 21 | 4 | | 22 | 11 | | 23 | 6 | | 24 | 10 | | 25 | 4 | | 26 | 6 | | 27 | 13 | | 28 | 6 | | 29 | 8 | | 30 | 6 | | 31 | 8 | | 32 | 17 | | 33 | 7 | | 34 | 4 | | 35 | 4 | | 36 | 11 | | 37 | 7 | | 38 | 9 | | 39 | 11 | | 40 | 4 | | 41 | 3 | | 42 | 4 | | 43 | 14 | | 44 | 19 | | 45 | 3 | | 46 | 4 | | 47 | 5 | | 48 | 3 | | 49 | 2 |
| |
| 46.30% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 5 | | diversityRatio | 0.3333333333333333 | | totalSentences | 54 | | uniqueOpeners | 18 | |
| 0.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 0 | | totalSentences | 36 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 10 | | totalSentences | 36 | | matches | | 0 | "She turned left, the pavement" | | 1 | "She had followed this lead" | | 2 | "He didn’t look up as" | | 3 | "She could go back." | | 4 | "She could turn around." | | 5 | "It was a prison for" | | 6 | "She didn’t have to." | | 7 | "He didn’t need to." | | 8 | "She didn’t have to." | | 9 | "She took the first step." |
| | ratio | 0.278 | |
| 1.67% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 33 | | totalSentences | 36 | | matches | | 0 | "The neon glow of the" | | 1 | "Detective Harlow Quinn’s fingers dug" | | 2 | "Rain streaked her face like" | | 3 | "The suspect had vanished into" | | 4 | "She turned left, the pavement" | | 5 | "The streetlights cast long, skeletal" | | 6 | "Quinn’s breath came steady, her" | | 7 | "She had followed this lead" | | 8 | "A figure stepped into the" | | 9 | "A man, broad-shouldered, his dark" | | 10 | "He didn’t look up as" | | 11 | "Quinn didn’t lower her weapon." | | 12 | "The man—Tomás Herrera—kicked a loose" | | 13 | "The rain drummed harder." | | 14 | "Tomás exhaled through his nose," | | 15 | "Quinn’s grip tightened on her" | | 16 | "The words hung between them," | | 17 | "Quinn’s eyes flicked to the" | | 18 | "Bone tokens glinted in the" | | 19 | "She could go back." |
| | ratio | 0.917 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 36 | | matches | (empty) | | ratio | 0 | |
| 95.24% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 15 | | technicalSentenceCount | 1 | | matches | | 0 | "The air below was colder, thicker, the scent of damp stone and something older—something that made her skin prickle." |
| |
| 0.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 2 | | uselessAdditionCount | 1 | | matches | | 0 | "Tomás repeated, his voice sharp" |
| |
| 94.44% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 2 | | fancyCount | 1 | | fancyTags | | 0 | "Tomás repeated (repeat)" |
| | dialogueSentences | 18 | | tagDensity | 0.111 | | leniency | 0.222 | | rawRatio | 0.5 | | effectiveRatio | 0.111 | |