MCP Server

Reservly includes a Model Context Protocol (MCP) server that lets AI assistants like Claude, ChatGPT, and others interact with your booking system through natural language. Customers can check availability, browse services, and book appointments through AI conversations.

Available Tools

The MCP server exposes seven tools. Each maps directly to a public REST API endpoint, so the data returned is identical.

Tool	Description	REST Equivalent
`get_business_info`	Get business profile, hours, contact, and social links	`GET /info`
`search_services`	List available services with prices and durations	`GET /services`
`list_staff`	Get staff members, optionally filtered by service	`GET /staff`
`check_availability`	Find available time slots for a date and service	`GET /slots`
`list_events`	Browse upcoming events with dates and capacity	`GET /events`
`list_rentals`	Browse rental items with prices	`GET /rentals`
`book_appointment`	Create a new booking with customer details	`POST /book`

Claude Desktop Setup

To connect your Reservly business to Claude Desktop, add the following to your claude_desktop_config.json file. Replace YOUR-BUSINESS-SLUG with your actual business slug (found in Settings on your dashboard).

claude_desktop_config.json

{
  "mcpServers": {
    "reservly": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://reservly.io/api/mcp/YOUR-BUSINESS-SLUG/sse"
      ]
    }
  }
}

json

On macOS, the config file is located at ~/Library/Application Support/Claude/claude_desktop_config.json. On Windows, it is at %APPDATA%\Claude\claude_desktop_config.json.

HTTP Transport

The MCP server is available over HTTP at:

https://reservly.io/api/mcp/{slug}/sse

text

Any MCP-compatible client can connect to this URL. The server uses the Streamable HTTP transport in stateless mode — each request is independent, which means no session tracking is required and the server scales horizontally.

The protocol uses JSON-RPC 2.0 over POST requests. Supported methods are initialize, tools/list, tools/call, and ping.

Tool Parameters

Most tools only require your business slug, which is injected automatically by the MCP endpoint. The two tools that accept additional parameters are:

check_availability

Parameter	Type	Required	Description
`date`	string	Yes	Date to check (YYYY-MM-DD)
`service_id`	string	Yes	Comma-separated service UUIDs
`staff_id`	string	No	Filter slots by a specific staff member

book_appointment

Parameter	Type	Required	Description
`service_ids`	string[]	Yes	Array of service UUIDs to book
`date`	string	Yes	Booking date (YYYY-MM-DD)
`time`	string	Yes	Booking time (HH:MM, 24-hour)
`customer_name`	string	Yes	Customer full name
`customer_email`	string	Yes	Customer email address
`staff_id`	string	No	Preferred staff member
`customer_phone`	string	No	Customer phone number
`party_size`	number	No	Party size for group bookings (default 1)

Example Conversation

Here is how a typical AI-assisted booking conversation works. The AI assistant handles tool calls behind the scenes while the customer interacts in natural language.

Customer

What services does Luxe Salon offer?

AI calls search_services

AI Assistant

Luxe Salon offers Haircut ($45, 60 min), Blowout ($35, 45 min), Color Treatment ($120, 120 min), and several other services. Would you like to book one?

Customer

Book me a haircut for tomorrow at 2pm.

AI calls check_availability→ confirms 2:00 PM is open

AI calls book_appointment→ booking confirmed

AI Assistant

Done! Your haircut at Luxe Salon is booked for tomorrow at 2:00 PM. A confirmation has been sent to your email.