vapi-mcp

vapi-mcp

2
ai_chatbotcommunication

The Vapi MCP for Cursor project integrates Vapi's voice AI with Cursor, providing tools for voice calls, managing assistants, and conversation retrieval. It offers setup instructions, troubleshooting tips, and focuses on avoiding common configuration errors.

Vapi MCP for Cursor

This project implements a Model Context Protocol (MCP) server for integrating Vapi's voice AI capabilities with Cursor.

Setup Instructions

  • Project Structure: Contains TypeScript files, compiled JavaScript, and environment variables.

  • Environment Configuration: Requires specific Vapi API keys.

  • Building the Server: Instructions to build the server using npm.

  • Configuration in Cursor: Guidelines to avoid "Client Closed" errors.

Troubleshooting

"Client Closed" Error in Cursor

  • Check working directory and environment variables.
  • Ensure module type is set to ES modules.
  • Check permissions and test the server directly.

Module Not Found Errors

  • Verify correct directory and rebuild the project if needed.
  • Ensure all dependencies are installed.

Available Tools

  • vapi_call: Make outbound calls.
  • vapi_assistant: Manage voice assistants.
  • vapi_conversation: Retrieve conversation details.

Lessons Learned

  • Specify the working directory and environment variables explicitly.
  • Use absolute paths and "stdio" transport in Cursor MCP config.