mindbody-mcp

MCP server for Mindbody, the vertical CRM for wellness, fitness, and beauty studios (yoga, gyms, salons, spas, training studios). 5 tools for AI agents to fetch client info, query class schedules, book clients into classes (env-gated), book appointments (env-gated), and complete shopping cart checkout (payment-gated).

As of May 2026, Mindbody does not ship an official MCP server. Multiple community attempts exist but are unofficial and self-hosted. This is the production-quality, install-ready Python rail. Shipped same-day as Mindbody's May 28 2026 Bilt Rewards distribution partnership.

The 5 tools

Related MCP server: housecallpro-mcp

Install

pip install mindbody-mcp

Configure

export MINDBODY_API_KEY="your-mindbody-developer-api-key"
export MINDBODY_SITE_ID="your-mindbody-site-id-integer"   # use -99 for sandbox
export MINDBODY_USER_TOKEN="your-mindbody-staff-user-token"   # optional but required for client + write actions
export MINDBODY_ALLOW_BOOKINGS="false"   # set to "true" to enable booking writes
export MINDBODY_ALLOW_CHECKOUT="false"   # set to "true" to enable payment processing (separate flag for safety)

Get an API key in your Mindbody Developer account. The user token is obtained via staff login flow; check the Mindbody Public API V6 docs for details.

Use with Claude Desktop

{
  "mcpServers": {
    "mindbody": {
      "command": "mindbody-mcp",
      "env": {
        "MINDBODY_API_KEY": "your-mindbody-developer-api-key",
        "MINDBODY_SITE_ID": "your-mindbody-site-id-integer",
        "MINDBODY_USER_TOKEN": "your-mindbody-staff-user-token",
        "MINDBODY_ALLOW_BOOKINGS": "false",
        "MINDBODY_ALLOW_CHECKOUT": "false"
      }
    }
  }
}

Restart Claude Desktop. The 5 Mindbody tools are now available.

Use case: AI front-desk + AI personal scheduler

Read-only flow (safe, default):

1. Call get_client_info(client_id) to load the caller's profile

2. Call list_classes(start_datetime, end_datetime) to surface available classes

Booking-enabled flow (after setting MINDBODY_ALLOW_BOOKINGS=true):

3. Call book_client_to_class(client_id, class_id) to enroll the caller

4. Call book_appointment(client_id, staff_id, session_type_id, start_datetime) for 1-on-1 sessions

Payment-enabled flow (after setting MINDBODY_ALLOW_CHECKOUT=true):

5. Call checkout_shopping_cart(client_id, item_id, item_type, quantity) to process payment

Safety note: two separate flags for write actions and payment

• MINDBODY_ALLOW_BOOKINGS gates the two booking writes (class enrollment + appointment scheduling). Bookings create real reservations on the studio's calendar.

• MINDBODY_ALLOW_CHECKOUT gates the payment action. Checkout actually charges the client's payment method. This is intentionally a SEPARATE flag so studio owners can enable AI agent bookings without also enabling AI agent purchasing. Two-step opt-in keeps real money out of an over-eager agent's reach by default.

Billing note

Since October 1, 2023, Mindbody charges $0.002 per API call in production environments. Operators should cache responses where possible to control billing overhead. This MCP does not cache responses; caching is the operator's responsibility per workflow.

Architecture

• Public MIT-licensed wrapper around the Mindbody Public API V6

• Async HTTP via httpx

• pydantic v2 input validation

• Three-header auth: Api-Key, SiteId, optional Authorization (user token)

• Rate-limit aware (429 returns a clean error)

• Write actions gated by env flag

• Payment action gated by a SEPARATE env flag

Development

git clone https://github.com/NoBanks/mindbody-mcp.git
cd mindbody-mcp
pip install -e ".[dev]"
pytest

License

MIT. See LICENSE.

Author

Ryan Hammer (NoBanks). Solo founder + engineer. Built this and 15 other MCP servers as part of a sprint to expose AI agent rails for the products and platforms shipping daily. Built mindbody-mcp specifically as a same-day reply to Mindbody's May 28 2026 Bilt Rewards distribution partnership.

• GitHub: @NoBanks

• X/Twitter: @livingagentic

• Site: livingagentic.me, nohumannearby.com

Open to AI engineering roles, contract or full-time, remote-only.