Analyzing Narrative Prose
Analyze your writing against multi-lingual benchmarks through interactive visualizations, or inspect stylistic profiles, pacing charts, and similar works across 10,000+ indexed novels in English, Japanese, and Chinese.
The autumn rain lashed against the fog-stained windowpanes of 221B Baker Street as Inspector Lestrade threw open the heavy oak door. His greatcoat was drenched, and his eyes burned with feverish urgency. "Holmes, you must come at once," he gasped, steadying himself against the polished mahogany table. "Lord Harrington lies motionless in his locked study, with a shattered crystal decanter resting beside his chair." Sherlock Holmes did not rise immediately. He slowly set down his briar pipe, watching a slender ring of blue smoke drift toward the ceiling before adjusting his magnifying glass. "A locked room, Lestrade? How delightfully elementary. Tell me, did you detect the faint scent of bitter almonds upon the victim's lips?" Lestrade blinked in astonishment. "Why, yes, Holmes—how could you possibly know?" Holmes turned to me with a calm, razor-thin smile. "A classic case of cyanide poisoning, Watson. Pack your bag; the game is afoot."
Western frameworks often force plots into a rigid 3-act conflict formula. KishoLens tracks Eastern Kishōtenketsu structure—Introduction (起), Development (承), Twist (転), and Resolution (結)—quantizing sentence-level sentiment to reveal how narratives build momentum through thematic contrast, sudden shifts, and pacing harmony.
Every author carries an unconscious stylistic fingerprint. Using 384-dimensional dense semantic embeddings (all-MiniLM-L6-v2) and cosine similarity across 10,000+ indexed works, KishoLens discovers closely matching novels and maps where your prose sits between serialized web fiction and classical literature across 17 genre archetypes.
Grammar behaves differently in Japanese, Chinese, and English. We measure sentence clause depth, dialogue cadence, and vocabulary richness using dedicated NLP pipelines: spaCy for English and Chinese dependency parsing, SudachiPy and pkuseg for morphological segmentation, and VADER for emotional trajectory.