MCP

Mentiontone exposes a Model Context Protocol server so you can ask your AI assistant questions about your team's sentiment data in natural language: trends over time, platform breakdowns, recent mentions, and more.

Authentication

Use a Mentiontone API key (mt_api_…) as a Bearer token. Create one in Team settings → API keys.

Connecting Claude Desktop

Claude Desktop's local config only launches stdio servers, so a bearer-token HTTP server needs the mcp-remote bridge. Find your MCP server URL in Team settings → MCP, then add this to your claude_desktop_config.json under mcpServers and restart Claude:

{
  "mentiontone": {
    "command": "npx",
    "args": [
      "mcp-remote",
      "https://your-mcp-server-url",
      "--header",
      "Authorization:${AUTH_HEADER}"
    ],
    "env": {
      "AUTH_HEADER": "Bearer mt_api_your_key_here"
    }
  }
}

Available tools

Tool	What it returns
analytics_summary	Sentiment totals for a time range (last month, 6 months, all time). Supports per-alert filtering.
analytics_trend	12-period time series (weekly or monthly) suitable for charts. Returns positive, neutral, negative counts per period.
analytics_by_platform	Sentiment breakdown per source platform for the selected range.
list_alerts	All monitoring alerts with their name, terms, active platforms, and UUID (use the UUID to scope other tools).
recent_mentions	Latest mentions with content, sentiment label, platform, and URL. Filterable by sentiment or alert.