CrowdOS

npx add-mcp 'https://api.crowdos.ai/mcp'

The protocol Claude already speaks

Model Context Protocol is Anthropic's open standard for connecting AI agents to external services. Claude Desktop, Cursor, Cline, Zed, and a growing list of agent frameworks (LangGraph, CrewAI, AutoGPT, Devin) speak it natively. Drop an MCP server into the host's config and its tools show up in the model's tool picker — no glue code, no per-app SDK.

CrowdOS speaks it two ways: a hosted endpoint the API serves directly, and a pip-installable stdio server for desktop hosts that relays to it. Both offer the same tools for your account and handle request shaping, response trimming, and error sanitization.

LangGraph and CrewAI run the agent. CrowdOS is the crowd.

CrowdOS is the crowd an agent should query. LangGraph and CrewAI are hosts: the place the agent runs. Mount the server and the same tools show up in the model's tool picker. AutoGPT and Devin speak the protocol too, the same way Claude Desktop, Cursor and Cline do.

An agent role-playing a panel isn't enough

One model acting like N personas (effectively n=1).

An LLM persona is a costume for one prompt. Ours is a database row: census-grounded people with a fixed identity, an opinion history, and a standing news habit. Query that crowd. Do not ask one model to play the panel.

• Hosted: add a URL, nothing to install

  https://api.crowdos.ai/mcp
  

  The API serves the MCP endpoint itself, so it's always in sync with the platform — no package to update. Add it as a remote MCP server in claude.ai (Settings → Connectors), Cursor, or any client that supports remote servers. OAuth-capable clients walk you through sign-in and consent on first connect; the grant shows up as a revocable key under /account/api-keys. Clients that send headers instead use Authorization: Bearer crowd_... with an MCP key.

• Self-hosted: one pip, one config block

  pip install crowdos-mcp
  

  Then create an MCP key at /account/api-keys (every plan can) and paste this into your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

  {
    "mcpServers": {
      "crowdos": {
        "command": "crowdos-mcp",
        "env": {
          "CROWDOS_API_KEY": "crowd_..."
        }
      }
    }
  }
  

  Cursor and Cline use the same JSON shape in their MCP-server settings. Restart the host and the CrowdOS tools appear in the agent's tool picker.

The same tools, grouped by the question

The list below is every tool the server exposes. These four groups are the research jobs an agent is usually hired for.

• What do people think?

  run_focus_group is a voting study: a sentiment summary and representative quotes. run_debate is a multi-round debate: convergence and the shift from round to round.

• Which option wins?

  compare_options is an A/B/C/D test on text options: a Borda leaderboard and a winner. rank_items is MaxDiff: a rank order with bootstrap rank intervals.

• What would they pay?

  test_pricing runs a Van Westendorp or Gabor-Granger study: price points and a revenue curve.

• What will this study cost?

  preview_cost estimates the cost before the agent commits. It is free to call.

15 tools, agent-shaped

• Sentiment / voting study — sentiment_summary + representative quotes

• Multi-round simulated debate — convergence + per-round shifts

• A/B/C/D test on 2-10 text options (plan limits: free 3, pro 5, max 10) — Borda leaderboard + winner

• Van Westendorp / Gabor-Granger pricing study — price points + revenue curve

• MaxDiff on 4-200 items — rank order with bootstrap rank intervals

• Pretest 1-40 draft survey questions — verdict per question before you field it

• Preview how broad an audience a custom_audience filter describes — free

• Estimate credits + seconds before committing — agents budget here

• Discover audience templates

• Fetch full results of a previously-run sim

• Stop a running study and release the reservation

• Browse the live agent population (sanitized)

• Your saved panels (from persist_panel / evolve_panel runs) — metadata only, newest first

• One saved panel — size, expiry, demographic make-up, recorded sessions (never the roster)

• Delete a saved panel you own — immediate and permanent

preview_cost is the one most agents underuse — call it before any sim with population > 100 to see the credits and seconds estimate, then decide whether the budget allows. Free to call.

What an agent run looks like

> Run a focus group on whether companies should mandate
   4-day weeks. Use 200 agents from us_general_population.

[tool: preview_cost]   → estimate_credits=200 · ~33-88s
[tool: run_focus_group] → sim ad4b3736…
   sentiment_summary: { positive_pct: 71.5, negative_pct: 18.0,
                        neutral_pct: 10.5 }
   sample_responses: [ Maria Chen "It would be great for ...", ... ]
   billing.credits_charged: 200

The agent picks preview_cost because the population is non-trivial, sees that 200 credits is affordable, then commits to run_focus_group. The trimmed envelope (sentiment summary + 5 representative quotes + billing) fits in context; if the agent needs every agent's reasoning, get_simulation pulls the full payload.

The opinion benchmark, published

On the Pew opinion benchmark, the crowd's opinion distribution matches published polling at 93.3% parity, with a mean absolute error of 3.36 percentage points and a Pearson r of 0.984. That published figure is the opinion benchmark. Creative ranking and in-market ad performance are measured separately.

• Read the benchmark ledger
• How the crowd is built
• The methodology

On every plan

MCP works on every plan, Free included. Studies use credits from your balance, the same as on the web app — a small focus group uses about 20 credits.

Panels follow your plan, as on the web: up to 20 respondents on Free, 500 on Pro and 1,000 on Max, with 5, 30 and 120 requests a minute and a monthly MCP ceiling of 250, 25,000 and 125,000 credits. Pricing →

Every run_focus_group / run_debate response includes a billing.credits_charged field so the agent can keep its own running total.

The REST API is for Enterprise accounts

Calling CrowdOS from your own code rather than an agent host? The REST API is available to Enterprise accounts. Enterprise → LangGraph and CrewAI have community MCP-bridge packages that let you mount our server into a framework agent without running an MCP host.

The same crowd, in the browser

The MCP server is the way an agent runs a focus group. Pulse is a vote, a debate or a comparison. Screening Room pressure-tests a creative, a concept or a deck before you ship. Ads Labo reads which of two to four ad variants a panel prefers.

• Pulse
• Screening Room
• Ads Labo
• Pricing
• Synthetic focus groups
• Pre-launch testing
• Market validation

What is a focus group MCP server?

A focus group MCP server is a Model Context Protocol endpoint an AI agent can call. CrowdOS speaks it two ways: a hosted endpoint the API serves directly, and a pip-installable server for desktop hosts. Both offer the same tools for your account. Claude Desktop, Cursor, Cline, Zed, LangGraph, CrewAI, AutoGPT and Devin speak the protocol. Drop the server into the host's config and its tools show up in the model's tool picker.

Does it work with Claude, Cursor and Cline?

Yes. Claude Desktop, Cursor and Cline speak Model Context Protocol natively. Cursor and Cline use the same JSON shape in their MCP-server settings. Restart the host and the CrowdOS tools appear in the agent's tool picker.

How is CrowdOS different from LangGraph or CrewAI?

CrowdOS is the crowd an agent should query. LangGraph and CrewAI are hosts: the place the agent runs. LangGraph and CrewAI have community MCP-bridge packages that let you mount the CrowdOS server into a framework agent without running an MCP host.

Why not prompt an agent to role-play a panel?

One model acting like N personas (effectively n=1). An LLM persona is a costume for one prompt. Ours is a database row: census-grounded people with a fixed identity, an opinion history, and a standing news habit. Query that crowd. Do not ask one model to play the panel.

Can an AI agent run a focus group over MCP?

Yes. CrowdOS is a focus group MCP server for AI agents. On every plan, Free included, the agent calls run_focus_group and the study fields against a census-grounded panel that disagrees with itself. A small focus group costs about $0.40. Why not let one LLM role-play a panel? One model role-playing a panel is one voice. Query the crowd.

How accurate is the crowd an agent queries?

On the Pew opinion benchmark, the crowd's opinion distribution matches published polling at 93.3% parity, with a mean absolute error of 3.36 percentage points and a Pearson r of 0.984. That is the published opinion benchmark. The full ledger is on the benchmarks page. Creative ranking and in-market ad performance are measured separately.