WriteLike

Paste a passage to see how it is built: rhythm, clause density, word choice and tone, measured with spaCy and explained in plain language by Gemini.

95 words
⌘ / Ctrl + Enter
analysis.mdexample

Example output. Press Analyse this text to run it live.

Overall style & tone
Frequent lemmas like "generalize", "war", "true" and "death" suggest a serious, reflective, analytical tone. It leans formal rather than conversational.

Pacing — high oscillation in sentence and token length: concise statements mixed with longer, complex sentences create rhythm and emphasis.

Clarity — about 1.75 clauses per sentence keeps it digestible while retaining nuance.

To write like this

  • Vary sentence lengths for dynamic pacing.
  • Balance concise and complex vocabulary.
  • Explore abstract themes (truth, conflict, mortality).
  • Keep a serious, analytical tone.

English · Simplified Chinese (experimental). The backend runs on a free tier and may take a few seconds to wake up.

01What it does

WriteLike looks at the shape of a piece of writing (sentence length, rhythm, tone and word choice) and turns those patterns into feedback. It doesn't judge whether writing is good or bad; it shows how it works, so you can borrow a rhythm or a structure while keeping your own voice.

02How it works

The numbers describe what's happening; the language model explains what it means. Keeping the measurement deterministic makes the feedback explainable instead of a black-box opinion.

detect langdetect, English / Simplified Chinese parse spaCy 3.8 (en_core_web_sm, zh_core_web_sm), TextBlob sentiment measure sentence & token lengths, clause density, POS & dependency distributions, tense, lemma motifs explain compact JSON summary → Gemini 2.5 → style feedback serve FastAPI on Render · Next.js front end · HMAC-signed requests

03Posts about WriteLike