$ styleprofile demo --dir '<tmp>/bundled-demo' --no-syntax
[exit 0]
--- stdout
STYLE PROFILE   7 chunks, 4,496 words

As a reference   held-out Delta 0.68 typically, 95% under 0.89 (7 documents)
  Shorter texts: calibrated at 75 words (124 pieces), 150 words (56 pieces); not at 300 words (28 pieces from 7
  documents, worth 18 independent ones), as a length needs 20 independent pieces from 3 or more documents. Under 75
  words, no verdict.

Contrast: LLM drafts   5 chunks from 5 documents; separates them from the reference with AUC 1.00 (perfect separation on
7 reference and 5 contrast documents)
  LLM-likeness on held-out text: reference 0.20 typically (95% under 0.82), LLM drafts 14.23
  Length alone: AUC 0.73 (reference 640 words per chunk, LLM drafts 624): some length difference
  Compared with the reference, LLM drafts have: Long words (7+ letters) ▲, List items ▲, Bold phrases ▲, Headings ▲

Pass --all, or run `styleprofile show` on it, to see the metrics.

wrote <tmp>/bundled-demo/writer.profile.json
Next, score a draft against it:
  styleprofile score <draft> <tmp>/bundled-demo/writer.profile.json
STYLE COMPARISON   1 chunk, 542 words   vs writer.profile.json (7 chunks)

Overall: close   Delta 0.85
  Lower is closer. The reference's own held-out writing at this length (542 words) scores 0.71 typically (95% under
  0.99).

LLM-likeness: like the reference   0.71
  The reference's own writing at this length (542 words) scores 0.20 typically (95% under 0.86); the LLM drafts 14.20.

By area   Delta ÷ the top of the reference's usual range in each area; bars on a log scale
  close up to 1x, somewhat different to 1.5x, clearly different to 2x, very different above
  Voice and stance             1.32x  █████░░░░░░░░░░░░░░░  somewhat different
  Vocabulary                   0.72x  ███░░░░░░░░░░░░░░░░░  close
  Function words               0.66x  ███░░░░░░░░░░░░░░░░░  close
  Punctuation                  0.66x  ███░░░░░░░░░░░░░░░░░  close
  Sentence shape               0.49x  ██░░░░░░░░░░░░░░░░░░  close
  Rhythm                       0.47x  ██░░░░░░░░░░░░░░░░░░  close
  Formatting                   0.15x  █░░░░░░░░░░░░░░░░░░░  close

Biggest differences   (each ▲ or ▼ is one standard deviation, up to 3)
                                                   this text   reference
  Sentences per paragraph                               4.38        6.15   ▼▼▼
  "more"                                             5.5 /1k     1.5 /1k   ▲▲
  "but"                                              5.5 /1k     2.0 /1k   ▲▲
  Sentences starting However / Moreover                 2.9%        0.0%   ▲▲  reference never varies
  Em dashes                                          1.8 /1k     0.0 /1k   ▲▲  reference never varies
  LLM marker words (delve, crucial)                  1.8 /1k     0.0 /1k   ▲▲  reference never varies
  "these"                                            1.8 /1k     0.0 /1k   ▲▲  reference never varies
  Words per paragraph                             67.8 words  86.9 words   ▼▼

Pass --all to see every metric.

Next: try it on your own texts
  styleprofile build posts/ --contrast llm-drafts/ -o writer.json
  styleprofile score draft.md writer.json
  styleprofile score draft.md --against posts/
--- stderr
Thin reference: 7 chunks, 4,496 words (aim for 15+ chunks and 20,000+ words); add more of the writer's documents
