Get a real ballot on any round.
Most rounds are scored by lay judges who don't flow. Paste yours and get an objective ballot: who won, why, and speaker points per side. The AI reads the whole flow, names the key clashes, and writes the reason-for-decision the way a circuit judge would.
Reason for decision
Pro feedback
Con feedback
Built for the lay-judge gap
Public Forum runs on parent judges who never flowed a round in their life. You can give the cleanest summary of the year and still drop the ballot because the judge voted on who sounded more confident. This gives you an objective second opinion: a ballot that actually engages the arguments, walks the weighing, and tells you which clash decided the round.
Names the clashes
The RFD identifies the two or three key clashes by name, says who won each and why, and flags the arguments that got dropped. Drops decide rounds.
Speaker points per side
Both speakers get a score on the format's real scale, calibrated honestly. No clustering everyone at 28 to be polite.
Format-aware
It judges through the lens of your format: PF summary-to-final-focus rules, LD framework, Policy collapse, APDA tight calls. Not generic feedback.
How it works in three steps
One. Paste the full round. Both sides, every speech you have. Type the motion and pick the format. Two. Hit judge. The AI reads the whole flow and weighs the clashes the way a tournament judge does. Three. You get a winner, speaker points for each side, a reason-for-decision, and one concrete fix per speaker.
AI debate judge FAQ
Is there an AI that can judge a debate round?
Yes. Paste a full round here and the AI returns a tournament-grade ballot: winner, speaker points per side, and a reason-for-decision that names the key clashes and walks the weighing. It judges through your format's actual adjudication norms, not generic feedback.
Does this work for Public Forum?
Public Forum is the default format and the case this page was built for. PF runs on lay parent judges who don't flow, so an objective ballot that actually tracks the summary-to-final-focus chain and the first-versus-second-speaker rules is exactly the gap. LD, Policy, Congress, APDA, BP, Worlds, Asian Parliamentary, and a Quick Clash mode are all supported too.
How accurate is the AI ballot?
It runs the same adjudication engine as our live rounds: format-pinned judging rules, explicit weighing across magnitude, probability, timeframe, and reversibility, and a speaker-point scale calibrated to use the full range. Treat it as a sharp second opinion, not a replacement for a real tournament panel. It catches drops and weighing gaps that a lay judge misses.
What should I paste in?
The fuller the better. Both sides' speeches, labeled by speech if you can (Pro Constructive, Con Rebuttal, and so on). A clean transcript or a detailed flow both work. The more of the round the AI can see, the more specific the ballot.
Does it replace a tournament judge?
No. It gives you an objective ballot when you don't have a flowing judge in the room, and a second read after a round you think you got robbed in. The point is to fill the lay-judge gap, not to stand in for a real panel.
Is it free?
Yes, while DebateIt is in beta every tier is $0. Anonymous visitors get a handful of rounds; signing in with Google adds more. The post-beta plans on /pricing are reference only.