Use the Skool API as MCP tool for Claude / Cursor / Cline

Use case: You want your AI agent (Claude Desktop, Cursor, Cline, or any MCP client) to operate your Skool community directly — approve members, post content, audit threads — without writing custom integration code. Apify exposes any public actor as a Model Context Protocol server, so you get all 33 Skool API actions as discoverable AI tools with zero setup beyond config.

What is MCP?

Model Context Protocol is the standard that lets LLMs discover and call tools at runtime. Anthropic introduced it in late 2025 and the ecosystem grew fast in 2026 — Claude Desktop, Cursor, Cline, Continue, and most major AI dev tools support MCP servers out of the box.

Instead of writing a function-calling schema for each API you want your AI to use, you just point the AI at an MCP server URL and it auto-discovers what tools are available.

Why MCP for the Skool API?

Setup for Claude Desktop

1. Locate your Claude Desktop config

OS Path
macOS ~/Library/Application Support/Claude/claude_desktop_config.json
Windows %APPDATA%\Claude\claude_desktop_config.json
Linux ~/.config/Claude/claude_desktop_config.json

2. Add the Skool API MCP server

{
  "mcpServers": {
    "skool-api": {
      "command": "npx",
      "args": [
        "-y",
        "@apify/actors-mcp-server",
        "--actors",
        "cristiantala/skool-all-in-one-api"
      ],
      "env": {
        "APIFY_TOKEN": "your_apify_token_here"
      }
    }
  }
}

Get your Apify token from console.apify.com/account/integrations.

3. Restart Claude Desktop

The MCP server starts automatically when Claude launches. You’ll see a small hammer icon in the input box indicating tools are available.

4. Try it

In Claude Desktop, ask:

“Approve all pending members in my Skool community cagala-aprende-repite whose application mentions they’re founders. Reject the ones that look like spam.”

Claude will:

  1. Call members:pending to list applicants
  2. Review each whyJoin field
  3. Call members:approve or members:reject for each one
  4. Report back with summary

You’ll be prompted to confirm before each write action (Claude Desktop default behavior for MCP tools).

Setup for Cursor

Cursor uses the same MCP standard:

  1. Open Cursor Settings → Features → MCP Servers
  2. Click “Add new MCP server”
  3. Name: skool-api
  4. Type: stdio
  5. Command: 
    npx -y @apify/actors-mcp-server --actors cristiantala/skool-all-in-one-api
  6. Environment: APIFY_TOKEN=your_token
  7. Save and restart Cursor

Setup for Cline (VSCode extension)

  1. Open Cline panel in VSCode
  2. Settings icon → MCP Servers
  3. Add server with the same JSON config as Claude Desktop above
  4. Refresh Cline → tools appear

Alternative: Hosted MCP endpoint (no local install)

If you don’t want to install anything locally, Apify also exposes a remote MCP endpoint:

https://mcp.apify.com?tools=cristiantala/skool-all-in-one-api&token=YOUR_TOKEN

This works for any MCP client that supports HTTP transports (Anthropic’s web client, custom integrations).

Example prompts that work well

Audit your community:

“Find the 5 most active members in my-community based on posts in the last 30 days. Tell me who they are and what topics they post about.”

Bulk operations:

“Approve all pending members in my-community who joined in the last 24h.”

Content management:

“Create a post in my-community titled ‘Weekly check-in’ with this content: [text]. Pin it to the top.”

Course publishing:

“Take this markdown file [paste content] and publish it as a new lesson in the ‘Welcome’ course of my-community.”

Member outreach:

“List members of my-community who joined more than 30 days ago and haven’t posted yet. I want to send them a personalized DM (just show me the list, don’t message them).”

Caveats

Pricing

Same as any other Apify actor invocation. Each MCP tool call counts as one actor run with pay-per-event billing:

Typical “approve 10 pending members” run via MCP: ~$0.12.

More in Recipes