Do I need a RidgeHQ account to use this?

Yes. This is not a public or no-account service. You need a RidgeHQ account on the Grow or Scale plan — Starter plans don't include the AI assistant features. Ask us if you'd like to change your plan.

Which AI assistants can I connect?

Claude (web or desktop) and ChatGPT, using a custom connector over OAuth. Claude Code and other CLI or script-based MCP clients connect with a personal access token.

Can the assistant book, cancel, or refund anything?

No. It never creates or cancels a booking, takes a payment, or issues a refund, and it can't change settings, staff, or waiver templates. Those stay in the normal RidgeHQ screens.

What happens before a scheduling change actually runs?

For anything beyond a low-risk action, the assistant tells you exactly what it's about to do and waits for you to confirm. Nothing changes until you say yes, and the result is recorded in your audit log with an undo option where one applies.

How do I revoke access?

For a personal access token: Settings, Access tokens, Revoke — it stops working on its next request. For a connected assistant: disconnect it in Claude or ChatGPT, or deactivate the staff account that approved it.

Is my data isolated from other RidgeHQ accounts?

Yes. Every request is scoped to your own tenant and enforced at the database level, and it only sees what your staff role allows. The assistant never has more access than the person or token behind it.

Model Context Protocol (MCP)

Connect your AI assistant to RidgeHQ

Connect an AI assistant you already use — Claude, ChatGPT, or Claude Code — to your RidgeHQ account and operate day-to-day scheduling from the assistant. Ask it things like “What’s on today’s morning brief?” or “Who’s on Saturday’s boat?” and it answers from your live data. For everyday changes, such as moving a session, it asks you to confirm before it does anything.

Requires a RidgeHQ account on the Grow or Scale plan. This is not a public or no-account service — Starter plans don’t include the AI assistant features.

How to connect

An Owner or Manager sets this up in RidgeHQ under Settings, Access tokens, Connect an AI assistant. That page shows your server address (the MCP URL) with a copy button and the exact steps for your assistant.

Claude or ChatGPT (recommended)

No token to handle. Add a custom connector with the server address shown on the Connect page, then sign in with your own RidgeHQ staff account.

In your assistant, add a custom connector / custom MCP server.
Enter the server address from Settings, Access tokens.
Choose Connect, then sign in to RidgeHQ with your staff account.
On the consent screen, check the application name and the redirect address, then press Allow.

The assistant gets exactly your permissions — a platform administrator account can’t be used, because it belongs to no single business.

Claude Code & CLI clients

Create a personal access token (PAT) in Settings, Access tokens. Name it for the person and tool, and choose the lowest role that does the job.

Settings, Access tokens, Create token.
Name it (for example, “Sam, Claude Code”) and pick a role.
Copy the token — it’s shown once and can’t be shown again.
Add it to your client with an Authorization header, for example a one-line claude mcp add command for Claude Code.

Treat a token like a password: it is never put in chats, tickets, or shared documents. One token per person or tool makes it easy to revoke just one later.

What it can and can’t do

It can

Today's morning brief: alerts, trips today, unsigned waivers, gear and weather problems
Bookings, customers, and availability (sessions, gear, rooms)
Spots and conditions, waiver templates and compliance
Partners, commissions, staff, and the audit log
Check a participant in; show or hide a session on your booking page
Create a trip, link or unlink a session to it, suggest a weather reschedule
Assign, unassign, or move a participant between sessions (asks you to confirm first)
Create or reschedule a session, move a rental or accommodation block, block gear for maintenance, cancel a trip (asks you to confirm first)

Reads never change anything. Low-risk actions run straight away and are logged; everyday scheduling changes ask you to confirm first, every time.

It will never

Create or cancel a booking
Take a payment or issue a refund
Change settings, staff, partners, or waiver templates
Delete records
See another business's data — every request is locked to your own account

These stay in the normal RidgeHQ screens, by design, whatever role or token is used.

Security & privacy

Every request is scoped to one tenant; isolation is enforced at the database level on dedicated least-privilege roles, not just in application code.
Sign-in uses OAuth 2.1 with PKCE and a consent screen that always shows the exact redirect address, or a personal access token stored as a hash, never in plain text, and shown to you only once.
Every change made through an assistant is recorded in your audit log with the assistant as the origin, and most changes can be undone from there.
You can revoke a token or disconnect a connector at any time; access stops on the very next request.

Tool results pass through the assistant you chose, under that provider’s terms — only send what you’re comfortable sharing with them. See the full security page and privacy policy.

Listings & directories

RidgeHQ’s MCP server is being submitted to the official MCP Registry and to AI assistant connector directories. Links will appear here once those listings are live.

Frequently asked

Do I need a RidgeHQ account to use this?: Yes. This is not a public or no-account service. You need a RidgeHQ account on the Grow or Scale plan — Starter plans don't include the AI assistant features. Ask us if you'd like to change your plan.
Which AI assistants can I connect?: Claude (web or desktop) and ChatGPT, using a custom connector over OAuth. Claude Code and other CLI or script-based MCP clients connect with a personal access token.
Can the assistant book, cancel, or refund anything?: No. It never creates or cancels a booking, takes a payment, or issues a refund, and it can't change settings, staff, or waiver templates. Those stay in the normal RidgeHQ screens.
What happens before a scheduling change actually runs?: For anything beyond a low-risk action, the assistant tells you exactly what it's about to do and waits for you to confirm. Nothing changes until you say yes, and the result is recorded in your audit log with an undo option where one applies.
How do I revoke access?: For a personal access token: Settings, Access tokens, Revoke — it stops working on its next request. For a connected assistant: disconnect it in Claude or ChatGPT, or deactivate the staff account that approved it.
Is my data isolated from other RidgeHQ accounts?: Yes. Every request is scoped to your own tenant and enforced at the database level, and it only sees what your staff role allows. The assistant never has more access than the person or token behind it.