Speak money into existence.
Get matched, argue it out, and an AI judge picks the winner.
Real-time speaking practice
Choose a topic, speak against the clock, and get clear feedback when you finish.
Bet on your words
Debate a person or AI. Get a verdict at the end.
Speak out loud. It pushes back, tracks the clash, and writes your ballot.
Pick how you start
Post a motion, take a challenge, and get an AI judge ballot when the round ends.
The Samantha from Her, except it argues back. Speak out loud, get argued with, get a ballot.
Run rounds, onboard your school, help the next one start. Paid, part-time.
Post a motion, accept a challenge, and get an AI judge ballot when the round ends.
Speak against an AI opponent in a timed voice round. It pushes back, tracks clash, and writes the ballot.
Founding student ambassadors run rounds, onboard schools, and help the next program start. Paid, part-time.
Debatable.
Debate an AI or a real person in timed rounds. Get an instant judge ballot at the end.
still building. updating website every day.
Founding story + community Start debatingSpeaker points, winner, reason for decision. AI does this well, at a fraction of what a human judge costs.
The round turns on the clash between Pro's structural-harm framing and Con's network-value defense. Pro establishes three harms: measurable declines in adolescent mental health (NIH longitudinal data), algorithmic amplification of misinformation during election cycles, and the erosion of local community institutions as attention moves to platform feeds.
Con responds with the network case: crisis coordination in Ukraine, Iran, and Hong Kong; cross-border reach for diaspora and minority communities; small-business marketing access that traditional media never offered to anyone without a budget.
The judge weighs on three axes: reach, persistence, and counterfactual access. On reach, both sides land hits. On persistence, Pro wins clearly. Once attention infrastructure is captured, it doesn't return to local institutions. On counterfactual access, Con makes the strongest move of the round: the benefits have no real substitute (no print equivalent for a Tehran protestor), while many of the harms have substitutes (regulation, design changes, age-gating).
Pro wins because the mitigations Con leans on don't exist yet and the harms compound while we wait. The verdict is narrow. Pro's framing of 'structural' was largely conceded by Con, which is the only reason this isn't closer on the flow.
Today, the AI judge explains who won and why. Next, the people watching will vote on which side persuaded them. Across many rounds, those votes can show which arguments move Debatable participants. That is platform feedback, not a scientific poll of the wider public. See where rounds happen in the Debate Atlas.
The Polymarket of debate. Every round is a market on one question: which side will the AI judge vote for? Take a position while they are still speaking. The ballot settles it seconds after the round ends, and the earlier you called it, the more it moves your Sharp Score.
The market model, applied to rounds instead of world events. The AI ballot is what settles every market.
Play credits today. No cash value, nothing to buy, nothing to win.
Match with a real opponent on a real clock. The AI judge writes the ballot.
02Run a full-format round on the clock against an AI opponent that interrupts, takes POIs, and argues back. Finish with a written ballot.
03Drop in a speech or case. Winner, speaker points, and a written RFD.
04Build cases and rebuttals for any motion, in your format's structure.
05Formats, drills, and vocabulary, from first round to circuit level.
06Book a round for later, or watch one that is already running.
Other AI rooms: quick clash for one fast unstructured round, the full trainer for every format on real speech clocks.
One judged voice round becomes a public score, a percentile, and a link anyone can check. Free.
AI can polish anyone's essay. It cannot show you understood it. Answering a real person, on the clock, is the skill left.
AI already does this for youEssays, applications, cover letters, research memos.
It cannot hold your case out loud, against someone pushing back, on the clock.
La carga de la prueba sigue siendo suya, y no han dado ni un solo mecanismo.
The burden of proof is still theirs, and they have not given a single mechanism.
你说会有伤害，但没有说明它怎么发生。请给出机制。
You claim harm but never show how it happens. Give me the mechanism.
They argue in their language. You read yours. Translated live, both ways, in 20 languages.
Answer an open claim whenever you are ready, or join the community at the same time for weekly Spar Night.
Pick a claim. Record your side whenever you like. Your opponent answers when they are free.
AmaraWorld Schools · usually replies within a day
“Schools should ban phones during class.”
Claim posted by Amara
MeiPublic Forum · usually replies in a few hours
“Cities should make public transit free.”
Claim posted by Mei“AI art should qualify for copyright.”
Suggested for youChoose a debater, or use the AI topic pick.
Speak for 90 seconds whenever you are ready.
They answer laterYou are notified when their response arrives.
Answer back. The AI judges both sides.
Everyone joins at the same time. Enter the queue, get matched, debate live, and receive an AI result when the round ends.
Post a motion, take a side, and get a ballot when the clock runs out.
Find an opponentLive rounds against people and rounds against the AI both score. When the ballot comes back, your speaks land on the cross-format leaderboard. Best entry per debater, nothing else.
The leaderboard
Ranked across every format.
BP sits next to APDA sits next to PF. The ballot is written by the same judge either way, so the scale means the same thing in every row.
See the full leaderboard →Chat picks the motion, you argue it live, and the AI judge posts the ballot on screen. Half your chat will disagree with the call. That is the content.
Set up the stream judgeAlready streaming debate? See how it fits together.
Creators bring the audience. You bring the case. Enter an open bracket, clear the AI-seeded prelims, and earn the live matchup the whole stream gets to argue about.
Dream matchups only. No creator pictured is confirmed, affiliated, or endorsing this tournament.
Photos: TwitchCon 2019, Dale Cruse, CC BY 2.0; HasanAbi, Poisonwithahawkseye, CC BY 4.0; xQc, Esfand, CC BY 3.0; Pokimane, U.S. Congress, public domain.
Current motions across education, free speech, technology, rights, and public life.
Schools
Schools should require parental consent before using a student’s new name or pronouns.Core clash Parental authority vs. student privacy
Universities
Universities should remain neutral on political controversies.Core clash Public trust vs. moral leadership
Online speech
Social platforms should ban anonymous political accounts.Core clash Accountability vs. dissent
Public memory
Museums should return artifacts acquired under colonial rule.Core clash Restitution vs. stewardship
Reproductive rights
The state should guarantee abortion access through the first trimester.Core clash Bodily autonomy vs. fetal rights
Youth and technology
Governments should ban social media for children under sixteen.Core clash Child safety vs. family choice
Classroom, Harrison Keely, CC BY 4.0; UC Berkeley protest, Dick Wheeler / The Daily Californian archives, MIT License; Social media apps, Theresepersonne, CC BY-SA 4.0; Returned artifacts, Spc. Barbara Ospina / U.S. Army, public domain; Ottawa march, Toronto Star / Toronto Public Library Archive, public domain; Child with smartphone, Andi Graf, CC0. Images are cropped for display.
The real gap between circuits isn't talent, it's access to quality practice. Debatable closes it. The AI argues back at full strength, and the ballot tells you exactly where the round was won or lost.
Proof that AI can sharpen critical thinking rather than replace it. You can tell how much thought went into every layer, from case generation to the objectivity of the judging.
In four years of British Parliamentary, I'd never seen a tool able to summarize, refute, and analyze high-level argumentation until I used Debatable. It would have made a real difference in my prep for international majors.
In a world where we can outsource our thinking to AI, we need more tools to keep us sharp. Debatable is the chess.com of debate.
Short answers. If something's not here, the last one routes to a real inbox.
Debatable is one place to practice, compete, and improve at debate.
Debate a real person live, argue out loud against the AI, or use Prep, Learn, Judge, and Coach between rounds. You can also schedule rounds and earn a verifiable certificate.
Fifteen formats. Fourteen languages. Every completed round can end with a written ballot: who won, why, speaker points, and what to drill next. Free in beta.
Free in beta. One round without an account, 10 signed in with Google. No card.
The numbers on /pricing are the post-beta plan, public so you know what is coming. School or team access: email aidandavidhollinger@gmail.com.
It argues, it does not fabricate. It presses soft claims, asks for warrants, and tells you when your logic collapses. It will never invent studies, cards, or tournament lore.
The harder problem is the opposite: it can sound too complete, missing the texture of a real round.
Most hand you a wall of generic text. No clock, no format, no judge.
This runs a live round: a motion, a side, a clock that does not pause. It takes POIs in WSDC, reads at Policy speed, stays impromptu in APDA, then writes a real ballot.
It still says wrong things. Catching them is the drill.
Fourteen. English, Spanish, French, Portuguese, Mandarin, Arabic, Hindi, and more. Switch from the language pill and the AI argues, takes POIs, and writes the ballot in whichever you pick.
Quick Clash. Three-minute rounds, no rules to memorize. You take a side, the AI takes the other, the clock runs.
When that feels easy, /learn covers each format with rules, sample topics, and how judges score.
Not for the first round. After that, sign in with Google to save ballots across devices and switch on style learning, where the next opponent is biased against your habits.
Google only. No passwords, no marketing list.
Yes. School programs run on a season license: $550 for the whole roster after beta, billed once a season. /schools has the longer pitch for coaches.
Running 50+ debaters? Email aidandavidhollinger@gmail.com.
Asian Parliamentary, WSDC, British Parliamentary, APDA, Public Forum, Lincoln-Douglas, Policy, Student Congress, MUN, plus Quick Clash for warmups.
Each runs its own prompt. WSDC and BP take POIs, LD goes framework-first, Policy spreads tagged cards, APDA stays impromptu.
No.
What would be cheating: running it mid-round at a live competition, or passing AI-written case copy off as your own original prep.
You pick. Quick Clash is three minutes. A full APDA round is about 45, BP about 60. Stop mid-round and it counts as practice.
Thirty minutes, on the record. Tell us what is working, what is broken, and what would make you stay.
Get round alerts, tournament sign-ups, and the product updates worth opening.
Watch your inbox for round updates, application openings, and tournament invites.
