MCP server & REST API · live

Run your Skool community
from any AI assistant

Helper connects your Skool community to Claude and any Model Context Protocol client. Manage posts, members, courses, the calendar, and DMs in plain language — or call the REST API directly from your own code.

Connect once with your Skool admin login · 36 tools exposed · no scraping setup

What it does

Everything you do in Skool — as tools an AI can call

Helper exposes your community's real actions over MCP. Point an assistant at it and ask in natural language, or wire the same actions into automations.

Posts & feed

Create, edit, comment, pin, and like posts across any connected community — the full feed, programmatically.

Members & CRM

Approve or remove join requests, ban abusers, read application answers, pull payment history, export the roster as CSV.

Classroom & courses

Build courses, folders, and lessons; write Markdown bodies; upload or embed video. Spin up a curriculum from a prompt.

Calendar

List, create, edit, and delete events — including daily/weekly/monthly recurrence with end conditions.

Direct messages

Find members, open a DM channel, and send messages in one call — onboarding nudges, replies, broadcasts.

Secure by design

OAuth 2.1 + PKCE, an encrypted credential vault, per-account rate limits, and a 30-day audit log of every action.

How it works

From login to live in three steps

Create an account & get an API key.

Register in seconds. Your key (sk_live_…) authenticates every call.

Connect a Skool community.

One call with your Skool admin login. Helper logs in, encrypts the session in a vault, and keeps it fresh — your password is never stored in plaintext.

Add the MCP server to your assistant — or call the API.

Drop the connector URL into Claude (or any MCP client), or POST to /v1/execute. Then just ask: “approve all pending members” or “post this week's update.”

MCP connector URL

Add this as a custom connector in Claude or any MCP client.

REST base URL

Authenticate with Authorization: Bearer sk_live_…

Made for agents

Built so an LLM can drive it unattended

Tool descriptions tell the model exactly what to call next, results come back as clean JSON, and every action is rate-limited and audited. The docs include copy-paste examples for MCP clients and raw HTTP.