Validated · 45 peer-reviewed studies

Get an early read
before you spend on real research.

AnthroSim runs realistic group conversations with grounded AI participants, so you can find the confusing questions and missing angles first. Then validate what matters with real people.

Request Early Access

Accepting researchers and industry teams for the first cohort.

Validated by peer-reviewed research
Political Analysis · 2023
Out of One, Many: Using Language Models to Simulate Human Samples
Argyle et al., Princeton & Stanford
LLMs conditioned on sociodemographic backstories reliably mirror the belief structures of U.S. political subgroups. This is the empirical basis for synthetic population simulation.
Nature Computational Science · 2025
Automated Social Science: LLMs as Social Simulacra
Cui et al.
156 psychology and management experiments replicated using LLM agents. Main effects held in most cases. The largest validation of LLM-as-participant methodology to date.
45 peer-reviewed papers · See the full validity argument →
What it produces

Concrete, structured output you can act on.

anthro-profile --count 1 --guidance "policy researcher"
Generated Profile
Dr. Sarah Chen
Age: 34
Occupation: Public Policy Researcher, Brookings Institution
Personality: Analytical, collaborative, evidence-driven
Political stance: Supports evidence-based climate policy; skeptical of partisan framing
Communication: Asks clarifying questions, cites sources, occasionally dry humor
Values: Scientific rigor, institutional trust, pragmatic compromise
anthro-chat --room climate-policy --subject "carbon pricing"
Simulated Conversation
Dr. Sarah Chen [14:03:21]
Recent EPA data shows a 23% emissions reduction in states with carbon pricing. The mechanism works when implementation is consistent.
Marcus Webb [14:03:45]
That ignores energy poverty effects. My constituents can't absorb another fuel cost increase without direct rebates built in.
Moderator [14:04:02]
Bringing in Dr. Okafor, who has studied distributional effects of carbon markets in lower-income regions...
47 participants · 83 min · Support 58% / Oppose 27% / Undecided 15% A directional read for what to test with real people. These numbers are not poll results.
Who uses AnthroSim
Research
Social science, behavioral research, AI behavior studies
Learn more →
Enterprise
Stress-test AI systems against diverse human behavior
Learn more →
Market Research
Early concept reads before you field a panel
Learn more →
Political
Pressure-test messaging before real constituents
Learn more →
Legal
Simulate jury deliberation and test case theories
Learn more →
Sensitive Research
Explore medical, privacy, and hard-to-recruit populations
Learn more →

Profile generation

Full internal profiles are much larger than the compact cards shown here, grounded in harmonized survey data plus environmental and professional context.

Conversation simulation

Run simple chats for speed or orchestrated simulations for stronger conversational dynamics. 10-500 participants.

Project materials

Give participants briefs, source packets, product specs, or case materials to read from and reference during discussion.

Voting & analytics

Personality-driven voting, context-aware decisions, moderator-announced results, full JSON export.

View all capabilities →

Request Early Access

We're onboarding a first cohort: academic and applied researchers, market research firms, enterprise AI and product teams, law firms, and policy organizations.

Early access participants receive preferred rates and priority onboarding.

No commitment. Pricing finalized at launch.