ShouldaSaid gives you a real opponent — one that answers your strongest point, pushes back, and doesn't let you off easy. Write your case, argue three rounds, and get an adjudication ballot on the World Schools rubric, not just a vague "good job."
Practice nowTake a side on today's motion — or write your own — and argue three rounds against an AI opponent.
Practice nowCreate a room, share the link, and argue a real person instead. Each of you gets your own ballot.
Debate a friendSet a motion for a whole class; every student argues it and gets graded individually.
Run a classEach round is graded on criteria adapted from the World Schools debate rubric:
Content (0–40) — the quality of the argument itself: reasoning, evidence, and how well you engage with the opposing case.
Strategy (0–20) — whether you identify what actually matters in the debate and structure your case around it.
Style — not scored. There's no delivery to judge in a written round, so we don't fake one.
Most people are worse at arguing under pressure than they think — not for lack of intelligence, but for lack of practice. Real arguments come with adrenaline, emotion, and time pressure. Simulating one in a low-stakes setting helps you:
The AI in ShouldaSaid doesn't just agree with you or serve up a canned counterargument. It takes the assigned side and actively looks for weaknesses in your reasoning. If your first speech is vague or unsupported, it presses harder. If you make a strong point, it looks for the edge case. It rebuts your last point first, then extends its own case — the way a real debater would.
That makes it a meaningfully better practice partner than asking a friend to "play devil's advocate." Friends know when to let you win. An AI opponent has no such constraints — and unlike a general-purpose chat assistant tuned to be agreeable, ShouldaSaid's opponent is specifically prompted to hold its side and press the actual argument you made.
If you're here because of an actual argument — not a debate topic — Redo a Fight lets you type out what happened and argue it again, with the AI playing the other person and the same judge grading the ballot.
How is this different from ChatGPT?
General assistants are tuned to be agreeable — they'll validate your position and avoid sustained confrontation. ShouldaSaid's opponent is built to argue back and hold the other side, and it's followed by an independent AI judge that grades the round on a fixed rubric. It's a purpose-built training tool, not a conversational assistant.
Can I replay the same motion multiple times?
Yes. Running the same motion more than once is one of the most effective ways to use the trainer — you'll usually see your ballot improve. Each round starts fresh, so the AI won't remember your previous attempt.
Does it remember previous sessions?
No, each round is independent — that keeps practice honest. If you're logged in, your past ballots are saved so you can look back at them, but the AI itself starts fresh every round.
Can I use it to prepare for a real debate or competition?
Yes, within the limits stated on our classroom page: it's a fast, consistent first pass on Content and Strategy, not a certified assessment and not a substitute for a human adjudicator or coach.
No sign-up. No download. Works on any device. Available in English, German, Arabic, and Spanish.
Try it now