Summer camps run two weeks in July and cost two to five thousand dollars. Then the practice room closes until next summer. Debatable runs every night of the year: voice rounds against a format-accurate AI, a judge ballot after every speech, fifteen formats. Free during beta.
A round needs an opponent, a clock, and a judge. Most students cannot get all three on a Tuesday.
This camp runs 24/7 from a browser. A written ballot every time.
Every format judged by its native paradigm.
Free during the public beta.
A practice round needs an opponent, a clock, a format, and a judge willing to write a real ballot. Most students can't assemble all four on a Tuesday in October. A camp that runs 24/7 fixes that.
The AI takes its side seriously, runs the format, takes points of information mid-speech, and pushes back on the weak link.
Every speech is timed the way it would be at a tournament. Prep runs in real time. POIs land in the protected window. The format is the format.
Every round ends with a structured judge ballot. Verdict, speaker points on the 25–30 scale, the key clash, your best line, the line you should have said, your critical drops, and one drill to work on next.
You can read every speech back, including the AI's. A coach can review it later. A flow becomes a study artifact, not a memory.
Use both. A camp is your two weeks of coaching in July. Debatable is your opponent the other fifty weeks.
No installer, no Zoom link, no cohort to chase.
Quick Clash if you have never debated, APDA or BP if you have. The AI builds a case, you respond, the judge ballot tells you where you actually stand.
The ballot calls out specific drops and weak warrants. Rerun the same motion from the opposite side, or pick a sibling motion, until the gap closes.
Same format, different opponent. Switch from the rigorous Surgeon persona to the rhetorical Closer or the cross-ex Prosecutor. Different opponents stress different parts of your game.
Capstone round of the week. Save it for your coach, send it to a teammate, or post it on the /spar waitlist to find a human partner for the next one.
Most AI tools speak one flavor of debate English at every format. These run by their own rules.
American Parliamentary, impromptu, no fake citations. The format the team built this thing on.
APDA practice →Four teams, extensions, whip, Sofia 2026 manual rules including mandatory POIs.
BP practice →Three speakers a side, definitional debates, sharper engagement on the framing fight.
Asian Parli →Mixed prep and impromptu, three speakers, fourth-speaker reply. The dominant global high-school format.
Worlds practice →Value-and-criterion frameworks, philosophical and ethical clash, both traditional and circuit registers.
LD practice →Lay-accessible, evidence-driven, two-on-two. Includes the 1st-versus-2nd speaker asymmetry that actually matters.
PF practice →Tagged-card delivery, spread-permissive, K-aff aware, current Arctic and Labor topics live.
Policy practice →Seven speech positions, precedence and recency for the presiding officer, current docket loaded.
Congress practice →GA, Crisis, JCC, and Specialized committees. Position-paper grading and bloc dynamics included.
MUN practice →A casual round for anyone who has never debated. Three minutes, no jargon, just talk and the AI talks back.
Try Quick Clash →Summer camps end in August. Tournaments are weekends, months apart. This covers the weeks between.
Run the released motion or the topic announcement against the AI. Try every position, find the strongest argument on each side, draft a frontline against the predicted opposition strategy. The AI does not get bored running the same motion six times.
Lost a round? Rerun the motion against the AI on the position you actually held. Compare your speeches to what the AI generates. Use the judge ballot to find the clashes you under-engaged.
Cross-examination, POIs, rebuttals, weighing, framework debates, theory shells. Each one isolatable as a focused drill. The 16 AI personas give you different opponent styles for each.
Most debaters specialize in one format. Picking up a second format usually means months without practice partners. With an always-on AI opponent that knows fifteen formats, the switching cost drops to one evening of drilling.
The bottleneck is rep volume, not coaching quality.
Free during beta. Future tiers, posted for reference: BYOK $1/mo, Individual $10/year, Team $50/year.
It is a sparring partner and a judge. You pick a format, get a motion, give a real timed speech out loud, the AI responds in format with points of information mid-speech, and a judge ballot lands when the round ends. Closer to a practice room with a strong opponent than a chat window.
Summer camps give you human coaches, cohort culture, and two weeks of structured drills for two to five thousand dollars. Debatable gives you a 24/7 opponent and a judge that explains its ballot, for free during the beta. They are complements, not substitutes. Use a camp for instruction in the summer, use Debatable for the rep volume every other week of the year.
Fifteen. The competitive core: APDA, British Parliamentary, Asian Parliamentary, World Schools (WUDC), Karl Popper, Lincoln-Douglas, Public Forum, Policy (CX), Congressional Debate, and Model UN, plus Quick Clash for first-timers, Viva for oral exams, and courtroom, negotiation, and pitch-defense rooms. Each format runs by its native rules. Policy gets tagged-card delivery, LD gets value and criterion, APDA stays impromptu with no fake citations.
For volume and consistency, yes. For human chemistry and live coaching, no. Most debaters lose more rounds to insufficient practice than to bad practice. The bottleneck is reps, not quality. An always-on opponent that holds you to format and writes a ballot every round closes it.
No. The AI is the opponent, the AI is the judge, and the structure of the round comes from the format you pick. A coach makes everything better, but the floor is one user with a phone and a quiet room.
Even better. Coaches get practice volume sized for a whole program, 1,500 requests a month across 50 seats on the Team tier, without doubling their own hours. Send your team here for the rep work and use coach time for strategy and drill design. The schools page has the full pitch for educators.
A national APDA champion at the University of Chicago. The voice banks, format rulebooks, and judge paradigms were written by someone who has actually stood at the podium in those formats, not a general-purpose engineering team.
No. The free anonymous and signed-in tiers will stay free forever, with caps on monthly request volume. The full-feature tiers (BYOK, Individual, Team) will turn on after beta exits. Future pricing is published on the pricing page for transparency.
Post on the waitlist to find a human partner once you have your reps in.
Spar waitlist →