Maverank MCP server

Let AI agents (Claude, or anything speaking the Model Context Protocol) read your workspace's AI-visibility data and trigger scans.

Connect

The MCP server is hosted — nothing to install. Point your agent at https://maverank.com/mcp/ and authenticate with a team API key (create one in Team settings → API keys; Studio plan and up).

Claude Code (one command):

claude mcp add --transport http maverank https://maverank.com/mcp/ \
  --header "Authorization: Bearer <team API key>"

Claude Desktop / Cursor / any MCP client (JSON config):

{
  "mcpServers": {
    "maverank": {
      "type": "http",
      "url": "https://maverank.com/mcp/",
      "headers": {
        "Authorization": "Bearer <team API key>"
      }
    }
  }
}

Keys with the project_read permission cover every read tool; trigger_scan additionally needs project_write (and the Agency plan).

Tools

Tool	What it returns	Key permission
`list_projects`	Tracked brands with latest visibility scores	read
`get_dashboard`	Score, KPIs with period deltas, per-engine breakdown, trend (`range`: 7/30/90)	read
`get_competitive`	Share-of-voice matrix + attribute comparison	read
`get_sources`	Citation leaderboard with presence flags	read
`get_readiness`	Readiness score, pillars and the fix list	read
`trigger_scan`	Starts an on-demand scan (Agency plan)	write

Example prompts for your agent

"Which of my clients dropped visibility this week, and what's the top readiness fix for each?" — the agent chains list_projects → get_dashboard → get_readiness and drafts the client email for you.

← REST API reference