Connect AI assistants like Claude Desktop, Cursor, Windsurf, and n8n to your Revuloop surveys using the Model Context Protocol (MCP).
Revuloop provides two ways to integrate with AI assistants:
@revuloop/mcp-server) that exposes 30+ tools for AI assistants to create surveys, collect responses, and analyze results.Install the MCP server to give your AI assistant direct access to Revuloop:
// Claude Desktop config
{
"mcpServers": {
"revuloop": {
"command": "npx",
"args": ["@revuloop/mcp-server"],
"env": {
"REVULOOP_API_KEY": "rlk_live_your_key_here"
}
}
}
}Requires a BUSINESS subscription for API access. See the npm package docs for Cursor, Windsurf, n8n, and Claude Code setup.
These endpoints return data optimized for LLM consumption — aggregated, formatted, and designed for single-call retrieval. They use standard rlk_ API key authentication.
/api/v1/mcp/dashboardReturns all surveys with response counts, completion rates, trends, alerts, and recent activity. Replaces multiple analytics calls.
Scope: surveys:read
/api/v1/mcp/surveys/:id/summaryComprehensive survey summary with performance metrics, per-question statistics, and health assessment.
Scope: surveys:read
/api/v1/mcp/surveys/:id/insightsDeep analytical insights including executive summary, key findings, segments, patterns, and recommendations.
Scope: analytics:read
/api/v1/mcp/surveys/:id/responses/recentReturns the most recent responses (default: 10, max: 25) formatted for AI assistant consumption.
Scope: responses:read | Query: ?limit=10
The MCP server package provides 30+ tools organized by category:
List, create, generate from prompts, get, update, launch, pause, close, and delete surveys.
List, get, submit, export (JSON/CSV/Excel), and delete responses.
Get quantitative analytics, overview, AI analytics, generate deep analysis, and poll job status.
Create sessions, respond to questions, complete sessions, get analytics, and browse transcripts.
List, create, update, test, and delete webhook subscriptions.