Run the argument before it costs you anything.
Commander, adversary, intelligence, diplomat, auditor. Give them a scenario and changing conditions, and approve every round before it runs.
Four steps, no configuration screen.
Pick the shape
War games, mystery room, crisis command, innovation tournament, civilization lab, or something you define.
Assign the roles
Any model can take any role. The adversary is the seat that earns its keep.
Set conditions
Rounds, injected events, and how strictly the fiction is held.
Approve the preflight
Nothing is sent until you have read what is about to run. Every subsequent round needs approval too.
Seats
3–8
Typical cost
About 12 credits per run
Best for
Strategy work, pre-mortems, and pressure-testing a plan you already like
What it does and does not do.
The value is not prediction. It is that a structured adversary surfaces the objection your team is too invested to make, and the auditor keeps a record of which assumptions were never tested. Both are things a real meeting struggles to do.
Simulations are fiction. They are bounded to prevent them being used to plan actual harm, and the boundary is not adjustable.
The other tools.
See it running.
The demo workspace shows this tool on sample data. No account, no card.
Open the demo