What the engine does
Compromise segments sentences, tags English grammar, extracts entities and noun phrases, and normalizes verbs locally in your browser.
Close reading without a verdict
Put two readings on the same page.
Compare what the language actually does: who and where it names, which actions move it, which concepts cluster, and where a motif returns. The evidence is local. The interpretation stays yours.
Load the original demo or paste two English passages.
Evidence, not judgment
Analyze two passages to reveal comparable textual evidence.
A words0
A sentences0
A actions0
B words0
B sentences0
B actions0
PASSAGE A
No evidence yet
PASSAGE B
No evidence yet
Sentence concordance
Select a motif to inspect every supporting sentence.
Counts are evidence dimensions, never a realism score or interpretation.
Counts become useful only when they lead you back to a sentence.
What rules add
Visible grounded and conceptual lexicons create inspectable comparison lenses. They are clues to follow, not a hidden model or a literary quality score.
What remains human
Verify extraction, read recurrence in context, compare counterexamples, and decide which claim the passage can honestly support.