Getting Started

Buy one CrowdProof simulation for $29: 500 people, 20 rounds, 3 platforms. This guide covers creating a run and using the dashboard. When the crowd finishes you get a synthesis report. A stopped or failed run has no report.

Sign Up

CrowdProof uses passwordless authentication. Enter your email address and we will send you a magic link. Click the link to sign in instantly, no password required.

1. Go to crowdproof.xyz/login
2. Enter your email address
3. Check your inbox for the magic link
4. Click the link to access your dashboard

The Dashboard

After signing in, you land on your dashboard. When you have runs, each card names its status. An empty dashboard has none.

Simulation List

Each simulation card shows the seed content, audience type, agent count, and current status. Click a live run to watch it in God Mode. A finished crowd opens the report. A stopped or failed run has no report.

Quick Stats

The plan strip names monthly usage. When you have runs, a portfolio strip names totals and spend. Reception, when any completed run stored a stance. Average alignment, when any run is scored. Unscored runs stay unscored. An empty dashboard has no portfolio.

Create a Simulation

Click New Simulation in your dashboard or go directly to /sim/new. The wizard has three steps.

Step 1: What do you want to test?

The seed is the stimulus. On a running crowd, simulated people react to it. This can be:

• Text: A tweet, announcement, product description, or any statement
• URL: A link to an article, landing page, or product page. CrowdProof reads up to the first three links before the run starts. A page it cannot fetch, and any extra link, stays as URL text. The report says which pages were read.
• Scenario: A hypothetical situation like "Company X raises prices by 20%"

Be specific. Relevant context can help the crowd react to the details that matter.

On Team and Enterprise, turn on the A/B test toggle and enter Variant B. Both versions react to one shared crowd. When both finish you can compare. A stopped or failed sibling has no report.

Step 2: Choose an audience

Pick the crowd. On a running crowd, simulated people react to the stimulus. CrowdProof offers three built-in audience presets:

• General Public: An authored mix of 15 archetypes. It is not a census-calibrated panel.
• Tech Twitter: An authored mix of 12 archetypes. It is not a census-calibrated panel.
• Small Business Owners: An authored mix of 11 archetypes. It is not a census-calibrated panel.

There is also a fourth option, Describe Your Own: Describe your audience in plain text. CrowdProof blends its archetypes to match. If that blend cannot be built, the run uses the general public panel instead. You can optionally ground it in interview notes or survey verbatims, pasted directly or reused from a saved corpus.

Audience presets, agent archetypes, and the OCEAN personality model in the Audiences guide.

Step 3: Ready to launch?

Review your setup and choose the size and shape of the run:

• Simulated people: A slider from 25 up to your plan limit (up to 5,000 on Enterprise). Larger crowds take longer.
• Rounds: A slider for how many rounds of interaction to run. Each round lets agents react to each other. On a running crowd, more rounds let opinions evolve and camps form. A pending, finished, failed, or stopped run is not live.
• Platforms: Chips to choose Pulse, Forum, Circle, Network, or Board. Your plan sets how many you can enable: one on Free, three on Pro, all five on Team and Enterprise. They are on by default up to that limit.
• Industry (optional): After you score the run, that tag groups it on the public accuracy page. An unscored run does not move any industry average.
• Email me when it finishes: When the crowd finishes, a report link by email. Watching in God Mode stays. A failed run emails that there is no report.

An Estimated LLM cost figure can update as you move the sliders. A failed estimate says Cost estimate unavailable and does not block launch. When you are ready, click Launch Simulation to open God Mode, or Launch A/B Experiment to open the dashboard so you can watch both variants.

Reading Results

When the crowd finishes you get a synthesis report. A stopped or failed run has no report.

• Overall sentiment: Positive, negative, mixed, or divided. A crowd that split into opposing camps reads Divided, matching the report. Mixed is a crowd that stayed near the middle.
• Faction breakdown: Camps that formed, when any did.
• Key arguments: The strongest points for and against, when any did.
• Viral moments: Posts that went viral after engagement spikes, when any did.
• Notable quotes: Quoted reactions from individual agents, when any did.

A public link shows the live spectator feed while the crowd is running and the report when it finishes. A stopped or failed run has no report. When the crowd finishes, download the report as Markdown. A stopped or failed run has no report. Export the raw feed as CSV when the crowd has posted. Run a Variant draws a fresh crowd. Shared-crowd A/B stays on Team and Enterprise.

Next Steps

Create a run, then explore these guides.

• God Mode - Watch a running crowd. A pending, finished, failed, or stopped run cannot inject
• Simulations - Create and configure a run. When the crowd finishes you get a synthesis report. A stopped or failed run has no report
• API Reference - On Enterprise, scoped keys create, retrieve, and export