MCP server

Run AgentChoice from your AI tools

AgentChoice runs a Model Context Protocol (MCP) server, so AI assistants and agents (Claude, Cursor, and other MCP clients) can check your AI visibility and run verdicts directly. Available on the Enterprise plan.

Endpoint & auth

https://agentchoice.app/api/mcp

JSON-RPC over HTTP. Authenticate with an AgentChoice API key as a Bearer token. Create one in Account → API keys.

Tools

check_visibilityDoes AI name a brand for a buyer query? Returns appearance rate, share of voice, average rank, and the competitors named instead. Fast.

run_verdictStart a full verdict comparing a site vs competitors. Returns a verdict id immediately; poll get_verdict until complete.

get_verdictFetch a verdict by id: status, AI visibility, who AI recommends, and the top fixes.

Connect Claude Desktop or Claude Code

Add this to your MCP config (e.g. claude_desktop_config.json). The mcp-remote bridge connects the hosted server to your client.

{
  "mcpServers": {
    "agentchoice": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote", "https://agentchoice.app/api/mcp",
        "--header", "Authorization: Bearer YOUR_API_KEY"
      ]
    }
  }
}

Connect Cursor (or any MCP client)

npx mcp-remote https://agentchoice.app/api/mcp \
  --header "Authorization: Bearer YOUR_API_KEY"

Then just ask

“Use AgentChoice to check if my brand is recommended for ‘best CRM for small business’, and tell me what to fix.”

“Run an AgentChoice verdict for mysite.com vs competitor.com and summarize the top 3 fixes.”

Verdicts run in the background; the assistant will call get_verdict until the report is ready, then summarize it.

Get your API key

Create a key in your account, then paste it into the config above.