---
title: "Ticket Fairy MCP Server | Connect Your AI Assistant"
description: "Give Claude Code, Cursor or any MCP client safe access to your live events, tickets and orders. Runs on your own machine, signs in as you, and by default stops to ask before anything destructive."
canonical: "https://www.ticketfairy.com/mcp-server"
---

Ticket Fairy MCP server

# Let your AI assistant run your events

You can ask Claude Code, Cursor or another MCP client about live sales, or hand it routine setup work, through the Model Context Protocol, the standard AI assistants use to work with a system directly. Routine changes run when the assistant calls them. Cancellations and deletions stop for your approval by default.

The CLI's MCP server runs on your own machine, signs in as you, and can only do what your Ticket Fairy role already allows. If you would rather install nothing, use the [hosted server](#hosted).

  [Connect your assistant](#setup) [See the guardrails](#safety)

You get it free with your account, included with the Ticket Fairy CLI, with no hosting, client secrets or separate login.

      Your AI client's config

```json
{
  "mcpServers": {
    "ticketfairy": {
      "command": "ticketfairy",
      "args": ["mcp", "--read-only"]
    }
  }
}
```

You need only this configuration once you install the CLI and sign in. Remove `--read-only` when you want the assistant to make changes as well as read.

## What you can ask for

Your assistant can use Ticket Fairy's own actions and live event data, so it can answer from your events instead of interpreting a dashboard screenshot.

### "How is Saturday selling?"

 You get the answer without opening four dashboard pages: your assistant reads live sales, check-in counts and order detail for the event you are talking about.

### "Set up next month's show like the last one"

 Your assistant can clone an event, add ticket types, price them and schedule the on-sale. Routine changes run when the assistant calls them. By default, destructive changes stop and ask you before they run.

### "Cancel this free order"

 Your assistant can inspect the order, show you what it found and ask for approval before it cancels the order.

### "Pull the numbers into my own report"

 You can put the figures straight into a spreadsheet, a summary or another tool, because your assistant works from structured data rather than a screenshot.

## Connect your AI client

You arrange no hosting, request no keys and expose nothing to the internet.

1. ### Install the CLI

   You need Node 20 or newer. The CLI includes the MCP server.
      Terminal   
   ```sh
   $ npm install -g ticketfairy
   ```
2. ### Sign in once

   Your browser opens for you to sign in. The server uses this sign-in, so you have nothing else to configure.
      Terminal   
   ```sh
   $ ticketfairy login
   ```
3. ### Add it to your AI client

   Paste the copied configuration into your client's MCP settings, then restart the client.
      Then ask  
   ```text
   "Which of my events are selling slowest this week?"
   "How many people have checked in tonight?"
   "Clone last month's show for the 14th."
   ```

On a server or in a shared runner, set `TICKET_FAIRY_API_KEY` to a Personal Access Token instead of signing in. You find out at once if the server starts with no credentials, instead of watching the assistant fail one question at a time.

## What the assistant can reach

Your assistant uses Ticket Fairy's own actions instead of interpreting a dashboard screenshot. For other endpoints, it can use the general-purpose `api` tool, with GET requests by default.

- `Context` Read, set or clear the active brand and event, and check both before making a change.
- `Brands` List and select your brands, use list\_brand\_geographics\_scopes to find the available series, venue, country and city filters, or use brand stats with get\_brand\_stats for whole-brand geographic and demographic insights. Geographic reports can be filtered by event, series, venue, country or city.
- `Events` List and read events and sales performance, plus create, update, publish, unpublish, postpone, disclose a date, cancel, reinstate, clone and delete.
- `Tickets` List ticket types, and create, update or delete them.
- `Orders` List and inspect orders, cancel a free order or approve an order awaiting review. Approval may capture an authorized card payment and issues the tickets.
- `Customers` Search your CRM by name, email or phone, then open a customer's full record.
- `Add-ons` List add-ons, and create, update or delete them.
- `Invitations` List invitations, and resend or cancel one.
- `Promotions` List promo codes, and create, update or disable them.
- `Check-in` Read live check-in stats and the tickets behind them.
- `Webhooks` List, read, create, update and delete webhooks.
- `Team` List team members, and invite, change or revoke access.
- `Tours` List and read tours and their members, and create, update or delete a tour.
- `Tables` List table types, and create or update them.
- `Messaging` List campaigns, and send one after you approve it in your compatible AI client.
- `Event Operating System` Run the production side of an event and a venue: Today, advancing, run of show, rota, incidents, risks, holds, artist offers, the draft line-up board (cards, first choices and backups, areas, notes and "Not for us" artists) and availability requests that ask agencies which artists are free. Emails and sign-offs ask for your approval in your compatible AI client on the CLI server; the hosted server leaves them to the dashboard.
- `Vault` Read-only: account, balance, transactions and transfers.
- `Assets` Upload an image or file and get back the URL to use in a change.
- `Auth` Report who the server is signed in as. The token value is never shown.
- `Direct API access` Call any endpoint that has no dedicated tool. GET-only unless you start the server with --enable-api-passthrough.

**Raw API writes are a separate choice.** Start the server with `--enable-api-passthrough` to add POST, PATCH, PUT and DELETE to the `api` tool. Raw API writes do not use the separate approval prompt provided for named destructive actions, and this flag cannot be combined with `--read-only`.

**Vault tools are invitation only.** They read Vault account details for brands with Vault access and never make changes.

## Decide how much it can do

You set the mode in the config file before the server starts, so the assistant cannot talk its way out of it partway through a conversation.

| How you start it | What the assistant can do | Detail |
| --- | --- | --- |
| `ticketfairy mcp --read-only` | Reads only  Where to start | Your assistant gets only the reading tools, so it can answer questions about your events and cannot change anything, whatever it is asked to do. |
| `ticketfairy mcp` | Reads and makes routine changes | Routine changes run when the assistant calls them. Before a destructive action runs, your compatible AI client asks you to approve it. The model cannot answer that prompt for you. |
| `ticketfairy mcp --allow-destructive` | Skip destructive-action prompts | You keep every named Ticket Fairy action but lose the separate prompt before every destructive action. Use it only when you accept assistant-initiated destructive actions. |

### Approve deletions and cancellations in your AI client

Before a deletion or cancellation runs, your compatible AI client asks you to approve it and waits for your answer. The model cannot answer. If your client cannot show MCP approval prompts, make that change in your Ticket Fairy dashboard.

Routine changes, such as adding a ticket type, run when the assistant calls them. Start the server in read-only mode to leave out every tool that makes changes.

Your usual permissions still apply in every mode. The server acts as you, with your role and your brands. Use the same access and review process your organization applies to work in the dashboard.

## The MCP server runs on your machine

Your AI client starts the server as a local process and talks to it over its own input and output. This server has no address on the internet for an assistant to connect to on its own, so you set up nothing on our side and nothing runs when you are not using it.

The server runs locally, but the event, order and customer data it returns goes to your AI client and may be sent to that client's model provider. Use a client and account your organization approves, and check its data-retention policy before connecting live event data.

Your default brand is stored in the CLI profile. The active event and any MCP-only context disappear when that process stops. Use a separate named profile for each workspace, and check the selected brand and event before any change.

When a request fails, read its message before retrying. Then ask your assistant to repeat the selected brand and event so you know the next call is working in the right place.

   [Install the CLI](https://www.npmjs.com/package/ticketfairy) [What else the CLI does](https://www.ticketfairy.com/cli)

## Hosted servers

If you would rather install nothing, connect to the MCP servers that Ticket Fairy hosts over Streamable HTTP. The server for your account acts for you after you sign in. The events and guides servers need no account or key, and they cannot see or change anyone's account. Hosted servers support MCP 2026-07-28 with per-request metadata and `server/discover`, alongside the existing initialization-based clients.

- `https://www.theticketfairy.com/agentic/organizer/mcp` Your assistant works on your own events. You can add this address, or add it from the official MCP registry, then sign in to Ticket Fairy in your browser. This server has the CLI server's tools for reading and changing your events, tickets, orders and team, and it can do only what your team role allows. With Spellbook, open your brand's spells, run their buttons and draft new ones. Refunds, deletions, sends and other actions that cannot be undone are not on this server: you do them in the Ticket Fairy dashboard.
- `https://www.ticketfairy.com/agentic/mcp` Your assistant can send people to buy tickets: it searches live public events and reads one event from its link. In clients that support MCP Apps, you see the results as event cards.
- `https://www.ticketfairy.com/docs/mcp` Your assistant answers questions about Ticket Fairy from the published steps, by searching and reading the how-to guides.

   Your AI client's config

```json
{
  "mcpServers": {
    "ticketfairy": {
      "type": "http",
      "url": "https://www.theticketfairy.com/agentic/organizer/mcp"
    },
    "ticketfairy-events": {
      "type": "http",
      "url": "https://www.ticketfairy.com/agentic/mcp"
    },
    "ticketfairy-docs": {
      "type": "http",
      "url": "https://www.ticketfairy.com/docs/mcp"
    }
  }
}
```

You can find each tool, the limits and the server cards in the [AI agent documentation](https://www.ticketfairy.com/developers/ai-agents#mcp), with every other way an agent connects to Ticket Fairy.

## Read your event in a focused App

Add the address for the task you want to do to your assistant. Sign in to Ticket Fairy, review the read-only access and approve the connection. Choose an event in the view, or enter its exact event ID.

- **Event review** `https://www.theticketfairy.com/api/agent/organizer/v1/apps/event-review/mcp` Review event details, dates, venue and ticket releases with your existing team access.
- **Sales and orders** `https://www.theticketfairy.com/api/agent/organizer/v1/apps/sales/mcp` Read native sales reports, browse filtered order pages and open an exact order.
- **Show day** `https://www.theticketfairy.com/api/agent/organizer/v1/apps/show-day/mcp` Read Now and Next by stage, the run of show and day sheets with your existing team access.

Event review shows your dates, venue and ticket releases. Sales and orders shows the native ticket-sales report, order pages filtered by status or unresolved disputes, and an exact order. Show day lets you choose a day and stage, read Now and Next or the run of show, and open a department, performer or stage day sheet. Open the linked dashboard page for advanced controls or changes.

### Which assistants can show these views?

A host that supports MCP Apps can show the interactive view. A text-only MCP client gets the same native tool answers. Each App has its own read-only OAuth connection and a small set of tools; connect each App separately. The existing event organizer server and CLI still provide the wider tool set. Team permissions apply to every read, and Show day uses your existing Event Management access. A view that cannot load offers refresh and dashboard links; changing the event clears the previous event's data.

### What do the sales figures and order pages count?

The ticket-sales report excludes comp orders and keeps the native currency and team-role visibility. Order pages are current filtered reads, not a sales total, export or complete text search. Changing the status or dispute filter starts at page one. You can open an order by its exact ID; it must belong to the selected event. These Apps do not issue refunds, send messages or change an event.

## Already using AI inside Ticket Fairy?

You can use Fai, the AI teammate in your dashboard, with your team and nothing to install, limited to your Ticket Fairy permissions. On Free, her help counts towards your limited Fai allowance. Eligible ticketing work on paid plans can be included without using usage; some answers and other work can use it. The MCP server connects the assistant you already work in to its own set of Ticket Fairy tools, on your machine and next to your other tools. Both use your Ticket Fairy permissions and live platform data.

   [Meet Fai](https://www.ticketfairy.com/event-ticketing/ticket-fairy-ai)

## Questions about the MCP server

### What is the Model Context Protocol, in plain terms?

It is the standard way an AI assistant plugs into a system it was not built with. Instead of the assistant guessing its way around a web page, it gets a named list of actions it can take and data it can read. Ticket Fairy publishes that set, so you can work on your events from any assistant that supports the protocol.

### Where does the server run? Do you host it?

The CLI server runs on your own computer, started by your AI client. It is there only while you are using it, and no address on the internet reaches it. If you would rather install nothing, Ticket Fairy also hosts a server for event organizers: add it from the official MCP registry and sign in to Ticket Fairy in your browser. Two more hosted servers only read: one searches public events and one answers questions from the guides. Neither can see your account.

### Can the AI do something I did not agree to?

Only within the mode you chose. The CLI server can use only the tools that mode registers. In read-only mode it cannot change anything. In default mode, routine changes run when the assistant calls them, while destructive actions stop and ask you in a compatible AI client. The sign-in also carries your role, so your assistant cannot reach a brand you cannot access.

### How does it sign in?

You sign in once with `ticketfairy login`, and the CLI server uses that sign-in. The hosted server asks you to sign in to Ticket Fairy in your browser instead. Either way, you need no separate account or client secret, and your assistant never holds a password.

### Does it remember my events between sessions?

The default brand stays in your active CLI profile. The active event and other MCP context last only while the server process runs. Check the selected brand and event before asking your assistant to make a change.

### How is this different from Fai, the AI in my dashboard?

You use Fai inside your Ticket Fairy dashboard, for everyone on your team and with nothing to install. On Free, her help counts towards your limited Fai allowance. Eligible ticketing work on paid plans can be included without using usage; some answers and other work can use it. The MCP server is for the assistant you already work in, either on your own machine through the CLI, alongside your code and your other tools, or as the hosted server. Both use your Ticket Fairy permissions and live platform data.

See how the Ticket Fairy [event ticketing and registration platform](https://www.ticketfairy.com/event-ticketing/) helps you sell more tickets, with every feature, the pricing and the support you get.

## Further Reading

  [**AI agents** MCP servers, Agent2Agent, WebMCP, agent skills and discovery files.](https://www.ticketfairy.com/developers/ai-agents) [**Ticket Fairy CLI** Manage events from the terminal or a script.](https://www.ticketfairy.com/cli) [**Authentication** Personal Access Tokens, agent approval and OAuth.](https://www.ticketfairy.com/developers/authentication)

## Give your assistant live event data and controlled actions

Install the CLI, sign in once, and paste the configuration into your AI client. Start read-only, then enable tools that make changes when you are ready.

  [Install the CLI](https://www.npmjs.com/package/ticketfairy) [Create a free account](https://manage.ticketfairy.com/welcome) [Book a demo](https://www.ticketfairy.com/event-ticketing/book-a-demo)

Questions about connecting your assistant? Email [support@theticketfairy.com](mailto:support@theticketfairy.com).

The Ticket Fairy, Inc. is a fintech company and technology platform. The Ticket Fairy, Inc. is not a bank, financial institution, or regulated financial services provider. Vault banking services are provided through regulated banking partners.

---

This page is on Ticket Fairy, a ticketing, marketing and operations platform for live events and the venues that host them: festivals, club nights, concerts, conferences and tours, and night clubs, beach clubs and concert venues. It grows revenue with fan rewards that turn an audience into a sales channel, and runs the money side with payments and payouts, Vault business banking, and working capital advanced against ticket sales. [How Ticket Fairy ticketing works](https://www.ticketfairy.com/event-ticketing) · [Ticket Fairy for AI assistants](https://www.ticketfairy.com/llms.txt)
