MCP server

Pluto ships a hosted Model Context Protocol server, so an AI agent (Claude, Cursor, or your own) can call the whole Pluto API as tools (post to social, run an AEO scan, generate content, check usage) without you writing HTTP glue.

Endpoint

POST https://api.joinpluto.com/v1/mcp
Authorization: Bearer sk_live_...

Transport: HTTP (JSON-RPC 2.0); request maps to a single response.
Auth: your standard Pluto API key in the Authorization header. The same key, tier, quota, and scopes as the REST API apply, so MCP calls are metered identically.
Tools: 64 curated tools spanning social, ads, SEO/AEO, content, reviews, agents, and observability, plus auto-generated tools for the rest of the API.
Streaming (SSE) is on the roadmap; today each call returns a single response.

Methods

Standard MCP JSON-RPC methods:

tools/list: discover the available tools and their input schemas.
tools/call: invoke a tool with arguments.

curl https://api.joinpluto.com/v1/mcp \
  -H "Authorization: Bearer $PLUTO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Connect it to Claude

Add Pluto as a remote MCP server. In your MCP client config (Claude Desktop, Cursor, or any MCP-capable agent), point at the endpoint and pass your key:

{
  "mcpServers": {
    "pluto": {
      "url": "https://api.joinpluto.com/v1/mcp",
      "headers": { "Authorization": "Bearer sk_live_..." }
    }
  }
}

Once connected, ask the agent things like "run an AEO scan for Acme and tell me which platforms we're missing from" and it discovers the tool via tools/list and calls it with your key.

Notes

Use a read-scoped key if the agent should only read (usage, diagnostics, scans) and never post or spend.
Every MCP tool call shows up in your request logs (opens in a new tab) just like a REST call.
The full tool list mirrors the API Reference ↗; anything you can do over REST, you can do over MCP.

Authentication Webhooks