How does sayit's tempo and pacing analysis work?

sayit's tempo analysis measures whether your pacing was even by comparing every gap between your words against an ideal gap that changes depending on what punctuation, if any, comes before it — so a breath at the end of a sentence isn't judged the same way as a hesitation in the middle of a phrase.

30-second version: For each pair of adjacent words, sayit checks the actual silence between them against an ideal: about 0.06 seconds mid-phrase, 0.25 seconds after a comma or similar, 0.45 seconds at a sentence end. If the actual gap is meaningfully shorter than ideal, that stretch is marked rushed; meaningfully longer, dragging; within tolerance, good. The result rolls up into a Pace score with your words-per-minute and a plain-language read on your overall tendency — well-paced, rushing, over-pausing, or uneven.

Why does the "ideal" gap change depending on where you are in the sentence?

Because natural speech doesn't pause evenly. A fluent reader barely hesitates between two words in the middle of a clause, pauses briefly at a comma, and pauses more at the end of a sentence — that rhythm is part of what makes speech sound natural rather than robotic or rushed. Judging every gap against one flat ideal would penalize the pause that belongs at a sentence boundary exactly as hard as a hesitation that has no business being there. sayit reads the punctuation before each gap and applies the ideal that actually fits: roughly 0.06 seconds for a mid-phrase connection, 0.25 seconds after a comma, semicolon, colon, or dash, and 0.45 seconds at a full sentence break.

What counts as "rushed" versus "dragging"?

sayit compares your actual gap to that boundary's ideal, with some tolerance built in (about 0.12 seconds either way) before it calls anything out. Come in meaningfully under the ideal and the gap reads as rushed — you ran two words together where a listener would expect a small separation. Come in meaningfully over and it reads as dragging — you held a pause longer than the moment called for. If Whisper detects your next word actually starting before the previous one finished, that's clamped to a strong rushed rating outright, since overlapping speech is never the target.

What's the difference between the Pace card and the Fluency score?

They're related but not identical. Pace is specifically about the shape of your pauses — gap by gap, against the ideal for where each one fell — and reports your words-per-minute alongside a plain summary like "well-paced" or "rushing." Fluency is broader: it factors in your overall speech rate against a natural band, plus a penalty for how much of your read was silence and for filler-word density. The two numbers can move independently — you can pause in all the right places and still speak too fast overall, or vice versa.

What target rate is sayit coaching me toward?

130 words per minute is the coached target — a natural conversational pace, not a race. sayit also reports your raw words-per-minute alongside "articulation rate," which strips out pause time to show how fast you spoke during the parts where you were actually talking, separate from how much of the clip was silence.

Tempo at a glance

What does an "uneven" tendency mean?

It means your gaps don't follow a consistent pattern in either direction — some rushed, some dragging, without a single dominant tendency. That's a different fix from a consistently-fast or consistently-slow reader: it usually points at hesitation in specific spots (often where the text gets harder) rather than an overall speed problem, which is why sayit reports the tendency as its own category rather than folding everything into a single rushed-or-slow label.

Try it

Read a few sentences with real punctuation in sayit and check the Pace card in your results — it will show whether your gaps ran fast, slow, or matched the punctuation, next to your actual words-per-minute. The related pause and filler-word detection covers what happens with pauses that aren't tied to punctuation at all.