Watch opinions form, spread, and shift over time. A living simulation of how populations actually arrive at decisions.
Describe a population and a scenario. Get distributional predictions segmented by any attribute.
Simulate subscriber response to a password-sharing crackdown across 50M households. Model who upgrades, who churns, who pirates, and who discovers they never watched enough to care — segmented by household size, viewing hours, and price sensitivity.
Run a synthetic population through a 40% price increase on a subscription product. Surface the segments that absorb it, the ones that downgrade, and the vocal minority that drives cancellation narratives on social — before you ship.
Predict population-level response to military escalation in the Middle East. Agents form stances, have household conversations about budgets and safety, and self-organize into support, opposition, and disengagement as the crisis unfolds.
Model how a population adapts after a superintelligence demonstration. Track employment shifts, changes in sense of purpose, and daily routine disruption over months, segmented by career vulnerability, financial margin, and worldview.
Quantify the say-do gap. Identify which segments actually follow through on stated boycott intent versus perform outrage online, segmented by brand loyalty, social influence, and real switching costs.
Model how a new entrant disrupts an established category. Simulate trial, adoption, and switching behavior across customer segments — and predict which incumbent loyalists are actually persuadable versus locked in by habit, ecosystem, or contract.
You describe a population and a scenario. Extropy builds statistically grounded synthetic agents, connects them in a social network, and runs a multi-timestep simulation where each agent reasons, shares, and re-evaluates as opinions cascade through the network.
Describe who you want to simulate. An LLM discovers the attributes that actually matter for this population, researches real-world distributions with citations, and samples a statistically grounded set of synthetic agents.
Agents are connected through realistic social graphs. Who knows whom, who influences whom. The topology is derived from the population itself, not a generic template.
Each agent reasons individually through an LLM based on their persona, memory, and what they've heard from peers. This isn't a one-shot prediction. Agents re-reason as new information spreads, opinions shift, and convictions strengthen or erode across timesteps.
Distributional predictions segmented by any attribute. Broken down by income, geography, behavior, or whatever dimension matters. Track how sentiment and position distributions evolve across the full simulation timeline, not just the final state.
Every agent leaves a reasoning trace. Drill into any individual to see exactly what they heard, how they weighed it, and why they landed where they did. Understand the "why" behind every data point, not just the aggregate.
MIT licensed
Built for agent harnesses
OpenAI, Anthropic, or Azure
Same seed, same population