|
/ Documentation /For Developers/ LatePoint Abilities – Connect AI Clients to Your Booking System

LatePoint Abilities – Connect AI Clients to Your Booking System

LatePoint Abilities allow AI clients such as Claude Desktop, Claude Code, Cursor, Windsurf, and other Model Context Protocol (MCP) compatible applications to securely interact with your LatePoint booking system.

Instead of manually managing bookings from the WordPress dashboard, you can ask your AI assistant to perform tasks such as:

  • View appointments
  • Search customers
  • Create appointments
  • Update bookings
  • Manage services
  • View agents
  • Manage locations
  • View orders

LatePoint uses the WordPress Abilities API together with the Model Context Protocol (MCP), allowing compatible AI applications to securely communicate with your website.

Requirements

Before you begin, make sure you have:

  • WordPress 6.9 or later
  • LatePoint version with MCP support
  • WordPress MCP Adapter plugin installed and activated
  • Administrator access to your WordPress site
  • An MCP-compatible AI client, such as:
  • Claude Desktop
  • Claude Code
  • Cursor
  • Windsurf

Enabling LatePoint Abilities

mcp settings 1

Navigate to: LatePoint > Settings > General > MCP

You’ll see the MCP settings page.

Enable Abilities

Enable the Enable Abilities option.

This registers all available LatePoint abilities with the WordPress Abilities API.

By default, this provides read-only access.

Enable Edit Abilities (Optional)

Enable Edit Abilities if you want AI clients to create and modify data.

This allows AI assistants to:

  • Create appointments
  • Update appointments
  • Create customers
  • Update customers
  • Manage services
  • Manage agents
  • Manage orders

If disabled, AI assistants can only read data.

Enable Delete Abilities (Optional)

Enable Delete Abilities if you want AI assistants to permanently delete supported LatePoint records.

Warning Delete operations are permanent and should only be enabled if required.

Enable the MCP Server

mcp settings 2

Turn on Enable MCP Server.

This publishes the LatePoint MCP endpoint: https://yourwebsite.com/wp-json/latepoint/v1/mcp

Most users won’t need to manually use this endpoint, as LatePoint automatically generates the configuration required for supported AI clients.

Connect Your AI Client

After enabling the MCP Server, the MCP settings page displays configuration instructions.

  1. Select your AI client.
  2. Copy the generated configuration.
  3. Paste it into your AI client’s MCP configuration.
  4. Save the configuration.
  5. Restart your AI client.

Your AI assistant will now be able to access LatePoint according to the permissions you’ve enabled.

Available Permissions

SettingDescription
Enable AbilitiesAllows AI clients to read LatePoint data
Enable Edit AbilitiesAllows AI clients to create and update data
Enable Delete AbilitiesAllows AI clients to delete supported records
Enable MCP ServerMakes the MCP endpoint available to AI clients

Available Abilities

Depending on your permissions, AI assistants can perform actions such as:

Appointments

  • View appointments
  • Search appointments
  • Create appointments
  • Update appointments
  • Delete appointments

Customers

  • View customers
  • Search customers
  • Create customers
  • Update customers
  • Delete customers

Services

  • View services
  • Create services
  • Update services
  • Delete services

Agents

  • View agents
  • Create agents
  • Update agents
  • Delete agents

Locations

  • View locations
  • Create locations
  • Update locations
  • Delete locations

Orders

  • View orders
  • Update orders

The exact abilities available depend on the permissions you’ve enabled.

Recommended Configurations

Read-Only Access

Enable:

  • Enable Abilities
  • Enable MCP Server

Recommended if you only want your AI assistant to answer questions about your booking data.

Read & Edit Access

Enable:

  • Enable Abilities
  • Enable Edit Abilities
  • Enable MCP Server

Recommended if you want your AI assistant to manage appointments and customers.

Full Access

Enable:

  • Enable Abilities
  • Enable Edit Abilities
  • Enable Delete Abilities
  • Enable MCP Server

Only recommended if you trust your AI assistant with deleting data.

Security

LatePoint gives you complete control over what AI assistants can do.

  • Disable Edit Abilities to prevent modifications.
  • Disable Delete Abilities to prevent deletions.
  • Disable Enable Abilities to completely disable AI access.
  • Disable Enable MCP Server to stop all MCP connections.

Example Prompts

Once connected, you can ask your AI assistant questions like:

  • Show today’s appointments.
  • List tomorrow’s bookings.
  • Find customer John Smith.
  • Create an appointment for Sarah tomorrow at 3 PM.
  • Update John’s phone number.
  • List all available services.
  • Show all agents.
  • What appointments are scheduled this week?

Troubleshooting

MCP settings are not visible

Make sure you’re using:

  • WordPress 6.9 or later
  • A version of LatePoint that includes MCP support

AI client cannot connect

Verify that:

  • The WordPress MCP Adapter plugin is installed and activated.
  • The MCP Server is enabled.
  • The generated configuration has been copied correctly.
  • Your AI client supports MCP.

No abilities are available

Usually this means one of the following:

  • Enable Abilities is disabled.
  • The MCP Adapter plugin is inactive.
  • Your AI client needs to reconnect after changing permissions.

Frequently Asked Questions

Can I make my AI assistant read-only?

Yes. Enable Enable Abilities, but leave Edit Abilities and Delete Abilities disabled.

Can I disable AI access later?

Yes. Simply disable Enable Abilities or Enable MCP Server.

Which AI clients are supported?

Any AI client that supports the Model Context Protocol (MCP).

Can I control what my AI assistant is allowed to do?

Yes. LatePoint lets you independently enable:

  • Read access
  • Edit access
  • Delete access

This allows you to choose the level of access that best suits your workflow.

Was this doc helpful?
What went wrong?

We don't respond to the article feedback, we use it to improve our support content.