Skip to main content
Metaview supports the Model Context Protocol (MCP), an open standard for connecting AI tools to data sources. With MCP, you can connect AI-powered clients like Claude Desktop, Claude Code, Cursor, Windsurf, and others to your Metaview workspace and ask questions about your interview data in natural language. For example, you can ask your AI client to:
  • “Compare the candidates from my past 5 interviews”
  • “Find all the intro calls run by a specific interviewer”
  • “Give me a quick brief on a candidate before my next interview”
The AI client will query your sessions, pull summaries, and give you a structured answer.

​
Getting started

1

Navigate to the MCP settings

Go to Settings > MCP in the Metaview app, or navigate through the Integrations page.
2

Generate a new API key

Click the Create API Key button. Give your key a descriptive name (e.g. “Claude Desktop” or “Cursor”) so you can identify it later. Copy the generated key immediately — you won’t be able to see the full key again.
3

Configure your MCP client

Add the Metaview MCP server to your client’s configuration. See the client setup guides below for specific instructions.
4

Start using Metaview data in your AI tools

Once connected, your AI client can access your Metaview interview data, summaries, and reports through the MCP integration. Just ask it a question about your interviews.

​
Client setup guides

​
Claude Desktop

Claude Desktop does not yet support the "type": "http" format for remote MCP servers. To connect, you need Node.js installed and should use the mcp-remote bridge.
  1. Install Node.js if you don’t already have it.
  2. Open Claude Desktop and go to Settings > Developer > Edit Config.
  3. Add the following to your claude_desktop_config.json, replacing <api-key> with the key you generated in Metaview:
{
  "mcpServers": {
    "metaview": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.metaview.ai/mcp",
        "--header",
        "Authorization: Bearer <api-key>"
      ]
    }
  }
}
  1. Quit Claude Desktop and reopen it. You should see a tool icon indicating the Metaview MCP server is connected.

​
Claude Code

Claude Code supports remote MCP servers directly. Run the following command in your terminal, replacing <api-key> with your key: 
claude mcp add metaview --transport http https://mcp.metaview.ai/mcp --header "Authorization: Bearer <api-key>"

​
Cursor, Windsurf, and other MCP clients

Most MCP clients that support remote HTTP servers can use the following configuration. Add this to your client’s MCP server settings, replacing <api-key> with your key: 
{
  "metaview": {
    "type": "http",
    "url": "https://mcp.metaview.ai/mcp",
    "headers": {
      "Authorization": "Bearer <api-key>"
    }
  }
}
Refer to your client’s documentation for the exact location of MCP server configuration.
If your MCP client does not support remote HTTP servers, you can use the mcp-remote bridge approach described in the Claude Desktop section above. See the MCP documentation on connecting to remote servers for more details.

​
Managing API keys

You can create multiple API keys, one for each tool or client you want to connect. Each key is named so you can easily identify which tool it belongs to. To revoke access for a specific tool, delete the corresponding API key from the MCP settings page. This immediately prevents that tool from accessing your Metaview data.

​
Security

  • Each API key is scoped to your personal Metaview account
  • Keys can only access data you have permission to view
  • You can revoke any key at any time
  • API key previews are shown in the settings page so you can identify which key is which

​
Help and troubleshooting

If you have any issues connecting your MCP client to Metaview, please reach out to support@metaview.ai.