An MCP (Model Context Protocol) server for interacting with Skool.com communities.

Skool does not expose a public API, so this MCP uses authenticated requests to the Skool web app and internal API endpoints.

Current capabilities include:

• Reading community, post, member, notification, and Classroom data.
• Creating posts and comments.
• Replying to comments through Playwright when browser interaction is needed.
• Giving or removing a thumbUp/thumbs-up/like on posts and comments.
• Capturing screenshots of Skool posts.
• Discovering video sources on Skool pages, including Skool native/Mux metadata.

cd /path/to/skool-mcp
npm install
npm run build

By default the server looks for ~/.config/skool-mcp/config.json. You can override the path with the SKOOL_MCP_CONFIG environment variable:

export SKOOL_MCP_CONFIG=/custom/path/config.json

Create the config file at the chosen location.

Single-account format:

{
  "cookies": "auth_token=YOUR_JWT_HERE",
  "defaultCommunity": "sim-club",
  "baseUrl": "https://www.skool.com"
}

Multi-account format:

{
  "defaultAccount": "main",
  "baseUrl": "https://www.skool.com",
  "accounts": {
    "main": {
      "cookies": "auth_token=YOUR_JWT_HERE",
      "defaultCommunity": "sim-club"
    }
  }
}

To get cookies, open Skool in Chrome, then DevTools, Application, Cookies, and copy the auth_token value. The cookie string should be:

auth_token=YOUR_JWT_HERE

Add the server to your MCP client configuration:

{
  "mcpServers": {
    "skool": {
      "command": "node",
      "args": ["/path/to/skool-mcp/dist/index.js"],
      "env": {}
    }
  }
}

The MCP uses session cookies from an authenticated Skool session. Keep the config file private because the auth_token acts as your logged-in Skool session.

Most tools accept an optional account parameter. If omitted, the MCP uses defaultAccount in multi-account mode or the legacy top-level config.

Outbound Skool URL access is restricted.

Allowed hosts:

• https://www.skool.com
• https://api2.skool.com

The MCP blocks:

• Non-HTTPS URLs.
• Non-Skool hosts.
• URLs containing embedded username/password credentials.

This is especially important for skool_request and browser-backed tools that open a page URL.

• skool_request: Make an authenticated HTTP request to an allowed Skool URL. Useful for API discovery and debugging. Only www.skool.com and api2.skool.com are permitted.

• skool_community_info: Get community details, including name, description, member count, and settings.
• skool_community_labels: Get category labels for a community from api2.skool.com.

• skool_members_list: List community members with pagination.
• skool_members_pending: List pending membership requests.
• skool_members_approve: Approve a pending member. Requires admin/moderator access.
• skool_members_reject: Reject a pending member. Requires admin/moderator access.

• skool_posts_list: List posts in a community feed. Supports category, pagination, and sort order.
• skool_posts_get: Get a specific post with full content and comments.
• skool_posts_create: Create a new post in a community.
• skool_posts_comment: Add a top-level comment or reply to a comment using Playwright. This helps bypass WAF/browser-only flows.
• skool_posts_vote: Give or remove a thumbUp/thumbs-up/like on a post or comment.

skool_posts_vote is state-aware: it checks the current vote before acting so it does not accidentally toggle an existing like off.

Use:

{
  "postId": "POST_OR_COMMENT_UUID",
  "groupId": "GROUP_UUID",
  "vote": "up"
}

To remove your like:

{
  "postId": "POST_OR_COMMENT_UUID",
  "groupId": "GROUP_UUID",
  "vote": ""
}

• skool_post_screenshot: Capture a PNG screenshot of a Skool post using the configured auth cookies. Supports viewport screenshots and full-page screenshots.

Example:

{
  "community": "sim-club",
  "postSlug": "nike-x-river-plate-reto-sim30-semana-2",
  "fullPage": false,
  "viewportWidth": 1280,
  "viewportHeight": 900
}

• skool_video_discover: Discover video embeds and streaming candidates on a Skool page.

It can detect likely:

• Skool native/Mux videos.
• Loom embeds.
• Vimeo embeds.
• YouTube embeds.
• Wistia embeds.

The tool discovers sources only. It does not download videos.

By default, full media URLs are redacted because Skool/Mux URLs can contain temporary signed tokens.

Example:

{
  "url": "https://www.skool.com/sim-club/classroom/52c2b79b?md=be8edbb90c004ebdbc27b83ccb16196d",
  "pressPlay": true,
  "includeUrls": false,
  "waitMs": 5000
}

Typical result shape:

{
  "pageType": "classroom",
  "providerSummary": {
    "skool-native": 1
  },
  "videos": [
    {
      "provider": "skool-native",
      "sourceType": "metadata",
      "host": "stream.mux.com",
      "durationMs": 310766,
      "redactedUrl": "https://stream.mux.com/VIDEO_PLAYBACK_ID.m3u8"
    }
  ]
}

Use includeUrls: true only when you explicitly need full signed URLs and are comfortable handling temporary media tokens.

• skool_courses_list: List courses in a community Classroom.
• skool_lessons_list: List lessons/modules in a specific course.

• skool_notifications: Get recent notifications.

• Read operations mostly use Next.js data routes (/_next/data/{buildId}/...). The buildId is fetched dynamically.
• Write operations use api2.skool.com endpoints where available.
• Some comment and screenshot flows use Playwright because Skool behavior can be browser-dependent.
• A User-Agent header is required because CloudFront blocks bare requests with 403 responses.
• See API-DISCOVERY.md for endpoint notes and response shapes when available.