Toolbox · Hover diagnostics
Read the draft at full size and hover a marked sentence to find out what tripped it: passive construction with no actor, four hedges in a row, past thirty words. The diagnostic is anchored to the sentence itself, so the explanation never scrolls away from the evidence.
Twenty-five instruments on one bench.
Live tool
Offline · nothing leaves this tab
Loading the sentence diagnostics…
Side panels force a choice: read the draft or read the findings. A tooltip avoids it. The text stays at full measure with one dotted underline per problem, and the explanation appears only where you look, then leaves.
Floating UI does the part that usually breaks: collision detection. Near the top of the viewport the tooltip flips below the sentence, and near an edge it shifts inside, so a diagnostic is never cut off or hidden behind the line it describes.
Passive voice with no actor is the one that hides responsibility, and it is usually a decision rather than a slip. Sentence length past thirty words is a reading cost. Stacked hedges are a confidence problem: two qualifiers usually cancel each other.
Robotic openers are the fourth, and they are also counted in the panel as a share of every sentence in the draft, so a single flagged paragraph does not look like a pattern when it is not one.
Burstiness compares sentence lengths against their own average. Very flat prose has a low figure and tends to read as machine-generated, but plenty of competent technical writing is flat on purpose. Treat it as a hint about rhythm and fix it by varying sentence length where the argument allows, not by adding short sentences for their own sake.
Toolbox
The console runs structural, claim and formatting checks on the same draft, and a named reviewer takes on anything with consequences.
