MCP integration

MentionsAPI ships an MCP (Model Context Protocol) server so any compatible AI editor — Claude Code, Cursor, Windsurf — can call our API as native tools. Your AI assistant gets mentions_check, mentions_watch, mentions_discover, and mentions_compare as first-class tools.

What you get

  • mentions_check — your assistant can check brand mentions across 4 LLMs ($0.02) or scrape the Perplexity UI ($0.25). Mode picked automatically based on the natural-language ask.
  • mentions_watch — set up scheduled brand-rank watches with webhook delivery from inside the conversation.
  • mentions_discover — generate query candidates for a brand without leaving the editor.
  • mentions_compare — head-to-head delta between two brands or queries.

Quick setup

One-time install via npx. Replace lvk_live_... with your MentionsAPI key (create one at /app/keys).

bash
npx -y @mentionsapi/mcp@latest --api-key=lvk_live_...

The MCP server is a thin wrapper over our REST API — it forwards your bearer token and translates MCP JSON-RPC calls into HTTPS requests. No data is processed locally; your API key is the only secret.

Editor-specific setup

Pricing

Same as direct API access. Each MCP tool call is a single REST call billed at the mode's rate. The MCP server is free to install; the underlying API calls deduct from your wallet. Track spend at /v1/usage.