Getting Started

MCP Server Setup

Name: apitree
Author: apitree

Connect the apitree Remote MCP endpoint so AI Agents can discover and call APIs automatically via the Model Context Protocol — no install required.

Claude Code

bashclaude mcp add

claude mcp add --transport http apitree https://apitree.ai/api/mcp

With an API key (enables call_api, billing, forecasting):

bashclaude mcp add — with an API key

claude mcp add --transport http apitree https://apitree.ai/api/mcp \
  --header "Authorization: Bearer nxs_live_..."

Claude Desktop

Add to your claude_desktop_config.json:

jsonclaude_desktop_config.json

{
  "mcpServers": {
    "apitree": {
      "url": "https://apitree.ai/api/mcp",
      "headers": { "Authorization": "Bearer nxs_live_..." }
    }
  }
}

Cursor

Add to .cursor/mcp.json in your project:

json.cursor/mcp.json

{
  "mcpServers": {
    "apitree": {
      "url": "https://apitree.ai/api/mcp",
      "headers": { "Authorization": "Bearer nxs_live_..." }
    }
  }
}

The headers field is optional — discovery tools (search_apis, get_api_details, compare_apis, etc.) work without a key.

Local server (stdio)

Prefer a local process instead of the remote endpoint? Install @apitreedev/mcp-server via npx:

jsonMCP client config

{
  "mcpServers": {
    "apitree": {
      "command": "npx",
      "args": ["-y", "@apitreedev/mcp-server"],
      "env": { "APITREE_TOKEN": "nxs_live_..." }
    }
  }
}

Usage

Once connected, just ask your AI Agent naturally:

The agent uses search_apis and call_api tools behind the scenes.

promptAgent prompts

> "Find an API for weather data"
> "Call the IP geolocation API for 8.8.8.8"
> "What APIs are available for financial data?"