Questions people ask first.

You answer one question out loud, for about a minute. Sayforth draws your answer as a single line, tells you what happened, and gives you one change to carry into tomorrow. The whole thing takes about five minutes.

One horizontal line per take. A solid stroke is speech. A gap is a pause, exactly as wide as the pause was long. The heavier stretch is your main point. A small tick under the line is a hedge. Nothing else is drawn, so every mark means something.

Interviews are where most people start, because the date makes practice urgent. The same program works for meetings: updates, opinions, decisions. You choose Interview, Meetings, or Both on the first screen, and you can change it later.

Nothing bad. Practicing the next day repairs the missed one, once a week, and the days before still stand. On your Season page a repaired day shows as a small hollow mark. After two quiet days the app offers a 3-minute comeback session.

You answer the Day 1 question again, without hearing Day 1 first. Then you hear both, read both, and see both lines on one shared axis, with three measured numbers beside them. The evidence is your own two recordings.

No. There is no overall score or grade on any screen or in any notification. You see observations against your own Day 1. When a number has not really moved, Sayforth says it is about the same.

Sayforth does not score pronunciation or accent and will not. It measures pace, where pauses land, hedging language, and when your point arrives, which matter regardless of how you sound.

Because the count would not be honest. Speech recognizers often drop words like um before they reach the transcript, so any number we showed could be wrong. Sayforth only surfaces what it can measure reliably.

No. Pace, pauses, hedges, answer length, and where your point lands are measured on your iPhone. The only place a language model comes in is writing the short coaching note from numbers already computed. The product is the 30-day program and its structure.

No. Your recordings stay on your iPhone. We never upload your audio. Only the text of what you said is sent for coaching.

No. Your audio never leaves the phone, so it cannot be used to train anything. Transcript text sent for coaching is processed under OpenAI's API terms, and we request zero data retention where it is offered. Details are in the Privacy Policy.

Settings, then Delete all data. It removes recordings, transcripts, and progress from the phone and our backend. No account or email is required.

It is the real program: all 30 days, every question, unlimited takes with full on-device measurement, the Season page, the Day 30 comparison, and coached feedback on Days 1 to 3 and then one session a week. Daily coaching adds a coached read of every session and the Seasons after your first.

Yes, any time, from your iPhone: Settings, then Apple Account, then Subscriptions, then Sayforth. No call, no email, no retention flow. You keep access until the end of the period you paid for, and the free program keeps working after that.

The yearly plan starts with 7 days free, then $59.99 a year unless you cancel at least 24 hours before the week ends. We send a reminder 2 days ahead. Monthly is $9.99 starting today, with no trial.