Six frontier AIs debate.
One synthesizer answers.
Ask one hard question. Six frontier AIs argue it across 12 turns. You walk away with a layered record — a cited report, an advisory board review of that report, and the memo, proposal, or post written from both. The record is the product, not the transcript.
Two tiers, real models.
Two frontier models debate, with Perplexity web research feeding both agents shared, current facts whenever the question depends on them.
GPT-5, Gemini 3.1 Pro, Claude Sonnet 4.5, Grok 4, DeepSeek V3, Perplexity Sonar. We pick the roster for your question. File and image uploads allowed, deeper report.
Signup includes 10 free credits — one full Basic debate, no card. Membership is $20 per month for 250 credits, and members can add 400 more for $30 whenever they need them. Credits never expire.
A brief agent sharpens the question.
Most bad AI answers are bad questions. Before any model argues, a brief agent reads what you asked and — only when it genuinely changes the answer — asks up to three questions: who the decision is for, what outcome you need, what constraints are real. A clear question goes straight through without interrogation.
The resulting brief is shown at the top of your report, so you can see exactly what the six models were asked to settle.
No personas. Real difference.
We do not assign personas. Nobody is told to be "the skeptic" or "the optimist." Every agent gets the same system prompt.
The disagreement is emergent. Different training data, different RLHF, different reasoning styles — put them on the same ambiguous question and they naturally diverge. That's the entire thesis of the product.
Four constraints, nothing else.
You pick Decide, Stress-test, Research, or Compare. Every agent sees the same system prompt for that intent — e.g. Stress-test tells them to identify failure modes the others missed and reject vague generalities.
Before the first turn, Perplexity runs a neutral search on your question and injects a summary + sources into every agent's context. Nobody argues from stale training data.
Each turn is capped at ~180 words. This kills waffle, forces specificity, and keeps the round-trip fast.
Agents are instructed to cite prior turns by agent name when they disagree, and never to repeat themselves. Agreement is allowed — but it has to cost words, so it only happens when they really do agree.
Round-robin, full context.
A turn is one model response. A round is one complete pass in which every participating AI responds once. Basic runs 2 AIs for 6 turns (3 rounds); Premium runs 6 AIs for 12 turns (2 rounds). Agents take turns in a fixed rotation, and each turn sees the full transcript so far — every prior agent's argument, verbatim.
Round 1 · Turn 1 — Agent A opens. Round 1 · Turn 2 — Agent B reads A, responds, may push back. ... Round 1 · Turn 6 — every AI has spoken once. Round 1 complete. Round 2 · Turn 7 — Agent A reads all of Round 1, revises or defends. ... Round 2 · Turn 12 — Final word. → Synthesis Engine reads the whole transcript.
Not kumbaya. Not fighting for sport.
- →Frontier models from different labs really do disagree on ambiguous questions. That's the well-documented behavior — not a trick we engineer.
- →The intent prompt (especially Stress-test) explicitly rewards finding holes in the prior turn.
- →The word budget makes empty agreement expensive, so it only happens when it's real.
- →We do not assign personas. Nobody is told to be the skeptic or the optimist.
- →The Decide prompt says explicitly: update your view only when evidence warrants.
- →The synthesizer at the end rewards genuine convergence, so agents aren't punished for agreeing.
A separate model reads the transcript.
After the last round, Gemini 3.1 Pro reads the entire transcript with a separate synthesis prompt. It doesn't add new arguments — it distills what happened.
- · Verdict + confidence level
- · Points of agreement
- · Tensions, sorted strongest disagreement first
- · Cited sources
- · Minority Report — the strongest dissenting view, preserved
- · Executive summary
- · Risks with severity + mitigation
- · Unknowns that would sharpen the decision
- · Decision framework — the branching questions
- · Scenarios (base / best / worst)
- · Action plan (week 1, month 1, quarter 1)
The Minority Report is deliberate: when one agent stakes out a strong contrarian view, the synthesizer preserves it instead of averaging it away. Consensus is not always the truth.
A board review, then the document.
The report is the first layer, not the last. From the finished report you can run a Board Review: five decision frameworks — commercial economics, customer and reversibility, capital and downside, contrarian risk, and execution and market structure — re-read the verdict and either uphold it, modify it, or challenge it, each with reasons. They are frameworks, not impersonations of real people.
Then Make something turns the record into what you actually have to send: an executive memo, client proposal, stakeholder email, article, LinkedIn post, or landing page copy. A writer drafts it, a second model critiques it against the format's rules, and a revision pass fixes what it found. It writes only from what the report established — if the material for a format isn't there, it says so rather than inventing a number or a quote. Free with any report.
All three layers export together, in order, as one PDF or Markdown file. You can also share a read-only link and see when it's opened. See everything you get →
Debate Mode.
Today the mechanism is emergent disagreement — models disagree because they're different. A future Debate Mode will assign opposing sides, so you get structureddisagreement on questions that benefit from it. Both are useful; they answer different questions.