Students record a speech and upload it. The computer transcribes it and scores content, organisation, and evidence, alongside delivery pace and filler-word rate measured from the recording. You verify the ranking. No tournament, no judge pool, no travel.
No credit card • 30 students • 2 free contests every month
Create a Speech contest
Set the event, the time limit, and the deadline, and write your performance requirements into the contest rules — “performed standing, in front of an audience” — so every entrant sees them before they record.
Speakers record and upload video
A phone recording is enough. Entrants submit by link; no student accounts.
The computer transcribes, then scores
The speech is transcribed and scored on content, organisation, and evidence, alongside the delivery metrics measured from the recording: words per minute and filler-word rate.
You verify the round
Read the rationales, reorder anything you disagree with, then Confirm & Publish. Certificates and award pages follow.
Is this replacing human judges?
No — it is filling the rounds that currently go unjudged. You remain the judge of record and verify every result before it is published.
Are the videos used to train AI models?
Never. Video and transcripts are used to score that contest only, and never to train or improve any model.
What about consent for video?
Results start private, any student can be excluded from public display, and you can delete any submission at any time.
What kind of events does this fit?
Original-speech events first — oratory-style, informative, and prepared argument. Interp events built on published pieces are not the target today.
