Debate an AI opponent in a full timed round.
DebateAI.com runs format-accurate timed debate rounds against an AI opponent. Pick a motion, pick a side, and the AI argues the other side in real time. Speeches on the clock, points of information mid-speech, and a judge's Reason for Decision when the round ends.
Every major format, with its own register
- Asian Parliamentary · three-on-three with reply speeches, government/opposition split, accessible argumentation.
- World Schools (WSDC) · proposition/opposition, points of information, weighing across substantive and reply.
- British Parliamentary (BP) · four teams, extension speeches, whip collapse, no fabricated citations.
- APDA · American Parliamentary, impromptu, varsity college circuit; no rolling motion bank.
- Public Forum (PF) · two-on-two with crossfires, accessible to a lay judge, evidence with proper citations.
- Lincoln-Douglas (LD) · one-on-one, value and criterion framework, philosophy literature where the motion calls for it.
- Policy (CX) · tagged-card delivery, plan/counterplan/kritik, evidence on a clock.
- Student Congress · legislative format, authorship and floor speeches, rebuttal of prior speakers.
- Model UN (MUN) · committee speeches, resolution drafting, moderated and unmoderated caucus.
- Quick Clash · two-minute rapid format for warm-ups and casual rounds.
- Viva (oral exam) · academic defense mode; explain your reasoning out loud while the AI examiner pushes on the weak link.
What you get back when the round ends
A real judge's ballot. Not a vibe check. Example excerpt from a PF round on "USFG should substantially increase domestic semiconductor manufacturing":
Pro wins, 28.5 to 27.5 speaker points. Central clash was supply-chain resilience versus fiscal opportunity cost. Pro carried weighing on probability and timeframe: their CHIPS Act precedent argument was clean and gave specific 2024 build numbers (Arizona TSMC fab, Ohio Intel). Con's spending tradeoff was warranted but never weighed against the cost of a Taiwan strait disruption, which Pro extended in summary. Drop in final focus: Con's "private capital is sufficient" claim was rebuilt in their summary but the warrant (Morgan Stanley 2023 analyst note) was misrepresented per Pro's cross. To fix: collapse earlier in summary; signpost weighing before reading impacts; tighten the second contention into one analytic chain.
Built by a UChicago parliamentary debater
Most AI debate tools speak in generic "Harvard debate society" English. This one was built by someone who actually argued at the top of the American Parliamentary circuit. The format accuracy is the moat. An APDA round sounds nothing like a Policy round, and the AI knows that.