Usercall MCP
@junetic
MCP server that lets AI agents run real user interviews and retrieve themes and quotes.
Overview
What is Usercall MCP?
Usercall MCP lets AI agents run real user interviews via voice calls and return structured insights with themes and verbatim quotes. It integrates with MCP-compatible clients such as Claude Desktop and Cursor, enabling AI agents to gather qualitative user feedback autonomously.
How to use Usercall MCP?
Install via npm (npx -y @usercall/mcp), obtain an API key from app.usercall.co, and add the server configuration to your MCP client’s JSON settings (Claude Desktop or Cursor). Restart the client, then ask your agent to create a study, share the returned interview link with participants, and retrieve results using the provided tools.
Key features of Usercall MCP
- Runs real user interviews via voice calls.
- Returns structured insights with themes and verbatim quotes.
- Integrates with Claude Desktop, Cursor, and other MCP clients.
- Supports visual stimuli (images or Figma prototypes).
- Provides five tools: create, update, status, results, delete.
Use cases of Usercall MCP
- Understand why users drop off during onboarding by interviewing real users.
- Collect feedback on new design concepts (e.g., dashboard mockups or Figma prototypes).
- Automate user research to identify confusion points and friction in product flows.
- Gather qualitative insights without relying on synthetic feedback or assumptions.
FAQ from Usercall MCP
What is required to use Usercall MCP?
A valid Usercall API key and Node.js 18+.
How do I get a Usercall API key?
Sign in at app.usercall.co, go to Home → Developer → Create API key.
Can I show participants visual stimuli during interviews?
Yes, you can attach an image or Figma prototype URL when creating a study. Media is visible only to web participants; phone callers cannot see it.
What happens after I create a study?
Usercall MCP returns an interview_link that you can share with participants via email, Slack, Discord, or in-product prompts. The agent can later check the status and retrieve results.
How are results returned to the agent?
Once a study is complete, get_study_results returns structured themes with summaries and verbatim quotes. You can request a summary or full output.