---
url: /relyagent-whmcs.md
description: >-
  Turn your WHMCS into a self-hosted MCP server, exposing scope-gated WHMCS
  tools to AI clients like Claude, Cursor and ChatGPT over OAuth 2.0 — with
  per-connection tool allowlists, audit logging and rate limits.
---

# Relyagent for WHMCS

Relyagent turns your WHMCS install into a self-hosted **MCP server** — so AI assistants like
**Claude, Cursor and ChatGPT** can work with your WHMCS data through tools you control. You
decide which tools each connection may use, everyone signs in through your own WHMCS admin
login, and every call is logged.

![Relyagent dashboard in the WHMCS admin](/relyagent-whmcs/img/relyagent-index-dashboard.png)
*Addons → Relyagent → Dashboard*

::: info What is MCP?
The **Model Context Protocol (MCP)** is an open standard that lets AI assistants call external
tools. Relyagent makes your WHMCS one of those tool providers — safely, on your own server, with
no data sent anywhere except between your WHMCS and the AI client you connect.
:::

## Who it is for

A WHMCS operator who wants to *ask an AI assistant* to do real work against WHMCS — "summarise
this client's open tickets", "list overdue invoices", "draft a reply" — without building an
integration or handing out admin credentials. You expose a curated, read-only-by-default set of
tools and keep full control and an audit trail.

## How it works

Three ideas explain the whole product:

| Piece | What it is |
| --- | --- |
| **Tools** | 62 scope-gated WHMCS actions across 10 groups (clients, tickets, invoices, services, domains, admin, affiliates, email, quotes, orders) |
| **Connections** | An admin-provisioned link to one AI client, limited to the exact tools you tick, with its own OAuth credentials |
| **Guardrails** | Native WHMCS admin login + consent, per-call audit log, rate limits, prompt-injection sanitisation, and a dangerous-operations gate |

You create a connection, pick its tools, and hand the AI client its credentials. When it
connects, it discovers the OAuth endpoints, you sign in through your **normal WHMCS admin login**,
approve a **consent screen**, and it can then call only the tools you allowed.

## What you can do with it

* **Connect the big AI clients** — Claude Desktop, Claude.ai, Claude Code, Cursor and ChatGPT.
  See the [connection guides](/relyagent-whmcs/guides/claude-desktop).
* **Expose exactly the right tools** — tick tools per connection; scopes are derived
  automatically, and destructive tools stay off behind a
  [dangerous-operations gate](/relyagent-whmcs/features/mcp-tools#dangerous-operations).
* **Stay in control** — every connection needs a secret you issue and a native admin login;
  see [Connections & OAuth](/relyagent-whmcs/features/connections-oauth).
* **See everything** — one [audit-log](/relyagent-whmcs/features/audit-and-rate-limits) row per
  tool call, with CSV export, plus per-minute and per-day rate limits.
* **Trust the deployment** — [file-integrity](/relyagent-whmcs/features/file-integrity)
  verification and one-click [updates](/relyagent-whmcs/updating) with an automatic backup.

## Where to go next

* [Requirements](/relyagent-whmcs/requirements) — check your server first
* [Installation](/relyagent-whmcs/installation) — activate, license, enable connections
* [Configuration](/relyagent-whmcs/configuration) — create your first connection
* [MCP tools](/relyagent-whmcs/features/mcp-tools) — the 10 groups and the dangerous-op gate
* [Settings reference](/relyagent-whmcs/settings) · [Error reference](/relyagent-whmcs/errors)

::: tip New to Relyagent?
Work through [Requirements](/relyagent-whmcs/requirements) →
[Installation](/relyagent-whmcs/installation) → [Configuration](/relyagent-whmcs/configuration),
then open the [Setup guide](/relyagent-whmcs/guides/claude-desktop) for your AI client.
:::