The population is shown as a globe of thousands of points. Hovering or tapping a point zooms in; the point becomes a simple human figure that walks, calls, messages, talks, waits, or visits a place, and another agent reacts to it. Scrolling opens the globe into a ground plane where many agents act at once, then gathers them into the first report.
Watch behavior emerge.
Explore the simulation · Hover over an agent
Reports
Every run turns agent behavior into patterns, distributions, and the one finding that matters for the decision.
What do you want to simulate?
5,000 simulated agents · 14 simulated days
Key result
Purchase intent
after a 10% price increase
Behavioral response
Emergent insight
Price resistance rises sharply in the +8–12% range, especially among price-sensitive segments.
Decision support you can defend — with the honesty a serious decision deserves.
Every result is clearly labeled as a simulation — never presented as a real survey or a guarantee of how people will act.
Back-tested against 60 real, resolved public events — plus your own accuracy track record over time.
SOC 2 in progress · follows GDPR and standard DPAs · export or delete your data anytime · your inputs aren't used to train models.
Everything a first-time visitor wants to know.
New accounts get 100 one-time starter credits — enough to try the product across run modes. There is no recurring free allowance; for ongoing use, paid plans (Pro, Team, Enterprise) are coming.
No. Every result is produced by synthetic agents and is always framed as a simulation. It is decision support for exploring likely reactions — not a survey, and not a guarantee of how real people will act.
The engine is back-tested against 60 real, resolved public events, and you build your own accuracy track record over time. Treat the output as a directional, evidence-based signal rather than a precise prediction.
Minutes — a Standard run finishes in about 10 minutes and a Deep run in about 15. You can start a run, walk away, and get an email when it is ready.
Yes. Export a report as PDF, Word, or CSV, and share a read-only link with stakeholders — no account required to view it.
Your inputs are not used to train models, you can export or delete your data anytime, and we follow GDPR and DPA practices (SOC 2 is in progress). See the Trust center for details.
Professionals who make evidence-based decisions — individuals testing a product, price, message, or policy; boutique teams working in shared workspaces; and regulated enterprises. Typical uses span product and pricing, public opinion and policy, brand and messaging, customer segments, and workforce changes.
Standard (about 10 minutes) gives a fast, well-grounded read. Deep (about 15 minutes) adds a larger debate and deeper narrative analysis. Red Team re-runs the simulation under a library of adversarial stress scenarios and returns a go / no-go verdict — for high-stakes decisions you want to pressure-test.