9:41
Claude GPT-5 Gemini Qwen Grok
iOS · Launching soon

Five AIs argue
your toughest
decisions.

An iOS app that convenes Claude, GPT-5, Gemini, Grok, and Qwen as a five-member council. They debate your toughest decisions across three rounds — right in your pocket — and return a single, defensible verdict.

You're in. Watch your inbox.
scroll
Why War Table

One AI gives you an answer.
War Table gives you the answer.

Every model has blind spots. Claude is careful but cautious. GPT-5 is decisive but breezy. Gemini reaches for data; Grok sniffs for what's missing; Qwen punches holes for sport.

Ask one of them, you get one perspective shaped by one set of biases. Convene all five, force them to challenge each other across three rounds, and you get something none of them produces alone — a verdict you can actually defend.

ChatGPT One model. One answer. Take it or rephrase.
Poe / You.com Switch between models manually. Still one at a time.
Multi-LLM aggregators Five answers in parallel. No challenge. No synthesis.
War Table Five models argue. Three rounds. One synthesized verdict.
See it in action

The app, on glass.

From the question on your mind to a verdict you can defend — the full debate plays out on your iPhone in under a minute.

Ask War Table a question Opening positions Council debating
01

Ask

Type or speak the decision you're stuck on. The council assembles in seconds.

02

Opening positions

Each of the five takes their stance independently. No collusion, no consensus — five sharp angles.

03

The council debates

Three rounds. Each model sees the others, rebuts, refines. Watch the energy flow between seats live.

04

Get the verdict

A clear final recommendation with a confidence rating, plus the full reasoning trail from each member.

The Council

Five minds.
Five sharp angles.
One verdict.

Each model carries a permanent role on the council. The roles don't rotate — they're designed to clash. That's what makes the debate honest.

Engineered, not assembled

Every seat is the product of hundreds of iterations of prompt engineering, testing, and tuning — the roles are designed around what each model is genuinely best at. Claude's nuance makes it the Synthesizer. GPT-5's decisive execution makes it the Pragmatist. Gemini's live web search makes it the Analyst. Grok's blunt contrarian tuning makes it the Skeptic. Qwen's structured reasoning makes it the Architect. Each LLM is doing what it does best — by design.

Claude
Synthesizer

Finds the through-line. Reconciles competing positions into the clearest, most defensible recommendation.

GPT-5
Pragmatist

"What actually happens when we ship this Monday?" Bias toward execution, momentum, real-world tradeoffs.

Gemini
Analyst

Brings the data. Surfaces base rates, comparable cases, and the second-order consequences nobody else flagged.

Qwen
Architect

Reframes the question itself. Decomposes the problem, exposes hidden assumptions, breaks false binaries others didn't even see.

Grok
Skeptic

Hunts for what's missing. Distribution, incentives, who's left out — Grok refuses the easy answer.

A council in your pocket

For every call that actually matters.

Big decisions don't happen at a desk. They happen on a walk, in the car, lying awake at 2 a.m. War Table is an iOS app — five world-class models, ready in five seconds, wherever you are.

iPhone-first. One tap to convene the council. Watch the debate unfold live. Save verdicts to revisit. Share the reasoning trail, not just a one-line answer.
Career

"Should I take this job offer?"

Founders

"Is this startup idea worth quitting my job for?"

Strategy

"Ship the v2 redesign or hold for Q3?"

Relationships

"Is it time for the hard conversation?"

Money

"Buy the house now or wait 18 months?"

Anything urgent

"What am I missing about this contract?"

The protocol

Three rounds.
One verdict you can defend.

Not parallel polling. Not majority vote. A real adversarial debate where each round builds on the last — and the council either converges or shows you exactly where it doesn't.

ROUND 1
i.
Opening positions

Each of the five takes the question independently. No collusion, no awareness of the others — five distinct opening arguments.

ROUND 2
ii.
Rebuttals

Every model sees the others' opening positions and responds. They concede, attack, sharpen. The weakest arguments get exposed; the strongest get pressure-tested.

ROUND 3
iii.
The verdict

Refined final stances. The Synthesizer reconciles them into a single recommendation, with a confidence rating based on how much the council actually converged.

154/500
Founding Member spots claimed

Get a seat
at the table.

The first 500 waitlist members unlock launch-week pricing you can't get any other way. Lifetime access options reserved for Founding Members.

You're in. Watch your inbox.
Founding Member pricing Early TestFlight access Direct line to the founder