claude mcp add revuloop -- \
npx -y @revuloop/mcp-serverSet REVULOOP_API_KEY in your shell and restart.
Surveys your AI assistant
can actually use.
Paste one URL into Claude, Cursor or Windsurf and sign in. Your assistant can then build surveys, launch them and tell you what came back, in plain English, without leaving the chat. Nothing to install, no API key to create.
Paste the URL. Sign in. Done.
Revuloop runs a hosted MCP server, so there is nothing to install and no key to copy anywhere. Add the connector in your assistant, sign in to Revuloop once, and approve the connection. It works on every plan, including Free.
https://revuloop.com/api/mcpIn Claude, add it as a custom connector. In Cursor or Windsurf, add it as an MCP server URL. Your assistant opens a browser window, you sign in, and that is the whole setup.
Prefer to run it locally?
The same server ships as an npm package you can run yourself. It authenticates with a Revuloop API key, which is available on Starter and up. On Free, use the hosted connector above.
{
"mcpServers": {
"revuloop": {
"command": "npx",
"args": ["@revuloop/mcp-server"],
"env": {
"REVULOOP_API_KEY": "rlk_live_..."
}
}
}
}Add the MCP Client Tool node, point it at npx @revuloop/mcp-server, and connect it to an AI Agent node.
Your n8n workflow can now spin up surveys and analyze responses as part of any AI-agent pipeline.
Claude Desktop
Cursor
Windsurf
Claude Code
n8n
Coverage across the whole platform.
The Revuloop MCP server exposes the full surface area of the Revuloop platform as MCP tools, resources, and prompts, so your agent can go from idea to insight without leaving the conversation.
Surveys
Create, generate, launch, pause, update, or close surveys.
Responses
List, submit, export, or remove individual responses.
Analytics
Pull live response stats, themes, sentiment, and AI-generated insights.
Smart sessions
Spin up adaptive AI interviews and stream answers in conversational turns.
Webhooks
Wire Revuloop events into your stack with first-class webhook management.
LLM-optimized reads
Pre-aggregated dashboard, summary, and recent-response endpoints designed for agent context.
What your assistant can reach, by plan
Your assistant is only offered the tools your plan can actually use, so it never suggests something that would be refused.
| Capability | Free | Starter | Pro | Business |
|---|---|---|---|---|
| Build, launch and manage surveys | ||||
| Read responses and analytics | ||||
| Run adaptive Smart Survey sessions | ||||
| Ask questions about your results in plain English | ||||
| Send email invitations | ||||
| Schedule surveys to open and close | ||||
| Open and resolve CX cases | ||||
| Manage CX alerts and alert rules | ||||
| Manage webhooks | ||||
| Account health scores | ||||
| Organization-wide CX dashboard |
Built for AI assistants, not adapted for them.
Most survey platforms have a REST API. Revuloop has a full MCP layer, with pre-aggregated endpoints, agent-ready prompts, and conversational session primitives.
Build surveys from plain English
Ask Claude to generate an NPS survey about onboarding. The MCP server calls create_survey, generate_survey, and launch_survey, no forms, no clicking.
Analyze responses inline
Your assistant reads pre-aggregated, LLM-optimized endpoints. Get themes, patterns, and correlations surfaced as conversational output, not JSON you have to parse.
Run Smart Survey sessions
Agents can create conversational sessions, submit adaptive responses, and complete sessions end-to-end. Powered by the Revuloop 5-Whys engine.
Manage webhooks from chat
Your agent can list, create, test, and debug webhook subscriptions without ever opening the dashboard.
Agent-ready resources & prompts
MCP resources expose revuloop://surveys and analytics feeds. Pre-built prompts for create-survey, analyze-responses, and explore-smart-sessions give agents a reliable starting point.
LLM-optimized payloads
Dedicated /api/v1/mcp endpoints return pre-aggregated data so your assistant spends context on reasoning, not on parsing raw response arrays.
No other survey platform ships an MCP server.
We audited the major survey tools. None of them are installable in Claude Desktop, Cursor, or n8n today.
| Capability | Revuloop | Typeform | SurveyMonkey | Qualtrics | Tally |
|---|---|---|---|---|---|
| Hosted connector, no install or API key | |||||
| Claude Desktop install | |||||
| Cursor / Windsurf support | |||||
| n8n native MCP tool | |||||
| LLM-optimized endpoints | |||||
| Adaptive AI survey sessions | |||||
| Agent-ready prompts |
Questions developers ask us first.
What is the Revuloop MCP server?
The Revuloop MCP server implements the Model Context Protocol, which lets AI assistants like Claude, Claude Code, Cursor, Windsurf and n8n create surveys, collect responses, run smart-survey sessions and analyze results through natural-language prompts, without opening the Revuloop dashboard. Revuloop hosts it at https://revuloop.com/api/mcp, so you connect by pasting that URL into your assistant and signing in. The same server is also published as an npm package (@revuloop/mcp-server) if you would rather run it locally.
What is Model Context Protocol (MCP)?
Model Context Protocol is an open standard from Anthropic for giving AI assistants access to external tools, data sources, and prompts. Any MCP-compatible client can install the Revuloop MCP server and immediately gain survey, response, and analytics capabilities.
How do I connect Revuloop to Claude?
Add https://revuloop.com/api/mcp as a custom connector in Claude, then sign in to Revuloop when the browser window opens and approve the connection. That is the whole setup: nothing to install and no API key to create. It works on every plan, including Free. If you would rather run the server locally, add an mcpServer entry that runs `npx @revuloop/mcp-server` and set REVULOOP_API_KEY to a Revuloop API key, which is available on Starter and up.
Does Revuloop work with Cursor, Windsurf, and Claude Code?
Yes. Cursor and Windsurf both accept a hosted MCP server URL, so pointing them at https://revuloop.com/api/mcp and signing in is enough. Claude Code supports `claude mcp add revuloop -- npx -y @revuloop/mcp-server` if you prefer to run the package locally with an API key.
Can AI agents create Revuloop surveys programmatically?
Yes. The Revuloop MCP server covers survey lifecycle end-to-end (creating, AI-generating from a prompt, launching, pausing, and closing surveys), so agents can build, deploy, and manage surveys through conversation.
Do I need an API key to connect my AI assistant?
No. The hosted connector at https://revuloop.com/api/mcp signs you in through Revuloop directly, so no key is involved and it works on every plan including Free. An API key is only needed if you choose to run the npm package locally, and API keys are available on Starter and up. To create one, open Settings then API Keys in your dashboard; test keys use the rlk_test_ prefix and live keys use rlk_live_.
What is the difference between the Revuloop API and the Revuloop MCP server?
The Revuloop API is a REST interface for any programmatic integration. The Revuloop MCP server is a thin wrapper that exposes a curated subset of that API as MCP tools, resources, and prompts, with dedicated LLM-optimized endpoints that return pre-aggregated data so AI assistants spend context on reasoning, not parsing raw JSON.
Does Revuloop support running adaptive AI survey conversations through an agent?
Yes. The Revuloop MCP server includes smart-session tools: create_smart_session, respond_to_smart_question, complete_smart_session, list_smart_sessions, get_smart_session, so agents can conduct conversational surveys powered by the Revuloop adaptive AI follow-up engine.
Does the Revuloop MCP server work with LangChain, CrewAI, or other agent frameworks?
Yes. Any MCP-compatible runtime can use the Revuloop MCP server. Frameworks that support the MCP spec (including LangChain, CrewAI, LlamaIndex, and self-hosted agent stacks) can register Revuloop as a tool provider.
What does the Revuloop MCP server cost?
Nothing extra. Connecting an AI assistant is included on every Revuloop plan, including Free, and there is no separate add-on to buy. What your assistant can do scales with your plan: Free and Starter cover surveys, responses and analytics, Pro adds CX cases, alerts and webhooks, and Business adds account health and the organization-wide CX dashboard. Usage counts against your plan's normal response and AI token allowances.
Deeper technical reading: MCP Integration docs ยท API Reference
Your next agent ships with real feedback data.
Start free, paste the connector URL into your assistant, and have it running your first survey before your coffee gets cold.