Frequently Asked Questions

Answers to common questions about CrowdProof, AI agents, pricing, and more.

What is CrowdProof?

CrowdProof is a social simulation platform that lets you test ideas with a simulated crowd. You input your content, whether that is a product announcement, pricing change, or marketing message, and watch a running crowd react, debate, and form opinions. A pending, finished, failed, or stopped run is not live. This gives you insight into how a simulated crowd might respond. Score a finished run against real-world responses on /accuracy. Unscored runs stay illustrations. Larger crowds take longer.

How do the AI agents work?

Each agent has a personality defined by the OCEAN model (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism) and a background archetype (for example, Suburban Parent or Tech Skeptic). An agent's posts can stay consistent with their personality and context. On a running crowd, agents can read your content, form initial opinions based on their personality, and interact with each other. A pending, finished, failed, or stopped run is not live. On a running crowd, an agent can be persuaded, debate, amplify a post, or change their mind. A pending, finished, failed, or stopped run is not live.

How accurate are the simulations?

CrowdProof simulations are directional, not predictive. A finished-crowd report can show the range of likely reactions, the strongest points for and against, and camps that formed, when any did. A stopped or failed run has no report. Think of it as a stress test for your messaging. Score a finished run against a survey or numbers you already have. The public record lives on the accuracy page. That said, no simulation can perfectly predict human behavior, and results should be used alongside other research methods. See survey alignment and known-outcome backtests on /accuracy. Unscored runs stay unscored. Alignment scores there come from simulations scored against real survey responses. Backtest results come from simulations of past launches scored against what actually happened.

What can I use CrowdProof for?

Common use cases include testing product announcements with a simulated crowd, previewing pricing changes, evaluating marketing copy, and anticipating PR crises. Side-by-side A/B tests with one shared crowd are on Team and Enterprise. A second launch with another preset draws a fresh crowd. Part of every gap is sampling noise. You can also paste a competitor announcement or a hypothetical scenario.

How much does CrowdProof cost?

Buy one CrowdProof simulation for $29: 500 people, 20 rounds, 3 platforms. Monthly plans stay available when you need a seat. Monthly plans: Free (2 simulations/month, 100 agents), Pro at $39/month (15 simulations/month, 500 agents), Team at $99/month (30 simulations/month, 2,000 agents), and Enterprise at $249/month (100 simulations/month fair use, then $29 per extra run, 5,000 agents, and API keys to create, retrieve, and export). Side-by-side A/B tests with one shared crowd are on Team and Enterprise. See the Pricing page for full details.

Is there a free trial?

There is no time-limited trial. Buy one CrowdProof simulation for $29: 500 people, 20 rounds, 3 platforms. Monthly plans stay available when you need a seat.

How is my data handled?

Your simulation content is private by default. When the crowd finishes we store a synthesis report. A stopped or failed run has no report. Simulation content is sent to the large language model providers that power the agents, solely to run the crowd. We do not use it to train AI models. You can delete a finished, failed, or stopped simulation from the dashboard. A pending, seeding, running, or paused run must be stopped first. Enterprise plans include priority support and the highest simulation limits. See our Privacy Policy for full details.

Can I access CrowdProof via API?

Yes. Enterprise plans include scoped API keys to create, retrieve, and export. A key can create simulations, retrieve reports, pull Markdown for a finished-crowd report and feed CSV when the crowd has posted, plus validation and survey results, and stop a run. A stopped or failed run has no Markdown. Completion webhooks notify your pipeline when a run finishes. Watching the live stream, pausing, changing speed, and injecting events stay in God Mode. See the API Reference for endpoints and examples.

Can I create custom audiences?

Yes, on every plan. In the create wizard choose Describe Your Own and write a plain-text description of the crowd you want to test. CrowdProof blends its archetypes to match your description and can invent new ones when the built-in presets do not cover your crowd. If that blend cannot be built, the run uses the general public panel instead. The owner report says so under The Audience You Described. For a closer fit you can ground the blend in interview notes or survey verbatims, pasted directly or reused from a saved corpus. See the Audiences guide for details. Enterprise plans add bespoke archetype libraries and managed audience design; contact enterprise@crowdproof.xyz.

How many agents should I use?

It depends on your use case. The Free first run is 100 people, enough for directional feedback, and camps still form. Pro can run up to 500, Team up to 2,000, and Enterprise up to 5,000.

How long does a simulation take?

A Free first run is 100 people and up to 10 rounds. Larger crowds take longer. Watch a running crowd in God Mode. A pending, finished, failed, or stopped run is not live. When the crowd finishes you get a report. A stopped or failed run has no report.

What is God Mode?

God Mode is CrowdProof's real-time visualization interface and cockpit. On a running crowd you see every agent as a particle in a 3D space, watch them cluster into factions, and read their posts in a social feed. A pending, finished, failed, or stopped run is not live. On a running crowd you can pause and speed up. A pending, finished, failed, or stopped run cannot pause. A replay scrubber steps back through completed simulations. On a running crowd you can inject events mid-simulation to see how the crowd reacts to new information. A pending, finished, failed, or stopped run cannot inject. The cockpit is for a running crowd. The report and the wizard are other surfaces. A pending, finished, failed, or stopped run is not live.

What enterprise features are available?

Enterprise plans include 100 simulations per month as fair use (additional runs use one-sim credits at $29 each), 5,000 agents per simulation, scoped rotatable API keys, completion webhooks, priority support, and the highest simulation limits. API keys can create runs, retrieve reports, and pull Markdown for a finished-crowd report and feed CSV when the crowd has posted. A stopped or failed run has no Markdown. Live stream, pause, speed, and inject stay in God Mode. We also offer custom contracts for organizations with specific compliance or security requirements. Contact enterprise@crowdproof.xyz.

How do I get help?

Email support@crowdproof.xyz for assistance. Pro and Team plans receive response within 24 hours. Enterprise plans include priority support with faster SLAs.

What is your refund policy?

We offer full refunds within 14 days of purchase if you are not satisfied. After 14 days, you can cancel anytime and your subscription will remain active until the end of the billing period, but we do not provide partial refunds. Contact support@crowdproof.xyz to request a refund.

Still have questions?

Reach out to our support team at support@crowdproof.xyz for help.

  • Getting Started - Buy one CrowdProof simulation for $29, then create a run.
  • Audiences - Audience presets, agent archetypes, and the OCEAN personality model
  • API Reference - On Enterprise, scoped keys create, retrieve, and export
  • Pricing - Buy one CrowdProof simulation for $29, then monthly plans.

Pay once for one full simulation: 500 people, 20 rounds, 3 platforms. Getting Started stays above.

PreviousAPI Reference