Find the guide. Pick the guide below that matches the task. If none fits, search all guides: connect an MCP client to https://www.ticketfairy.com/docs/mcp (Streamable HTTP, no account or key) and call search_guides with the task in plain words.
Read it before you act. Call get_guide with the guide's slug. It returns who the guide is for, what to have ready, every step and the common questions. Follow those steps; do not guess the dashboard from memory.
Do the work, or hand it over. Each guide below says what an agent can do:
- Command line. With the organiser's permission, install the official CLI (npm install -g ticketfairy) and sign in with ticketfairy login, or use a Personal Access Token that the organiser stores in TICKET_FAIRY_API_KEY. Then check the account and pick the brand: ticketfairy auth status, ticketfairy brand list and ticketfairy brand use <BRAND_ID>. Add --output json --non-interactive to every command you run that does not already choose its own mode, such as --output csv or --interactive; the commands below leave them out to stay short.
- MCP. ticketfairy mcp --read-only starts the CLI's MCP server with read tools. Drop --read-only only when the organiser wants you to make changes.
- Fai. Fai is the assistant inside the organiser's dashboard. When a task is not on the command line, give the organiser the suggested prompt to paste into Fai.
- Dashboard. Otherwise, send the organiser the guide link and walk them through it.
Keep secrets out of the chat. Codes, passwords and tokens that the organiser receives go into the Ticket Fairy dashboard: the page that asks for them, or Fai there when the guide says so. They never go into your conversation with the organiser.
Keep actions within the organiser's approval. Use their explicit approval for this exact action, or show what you will create, change, send, refund or cancel and wait for their yes. Installing a skill or asking for a draft does not authorise sending messages. Use only the approved brand, event and recipients, and only what the person's team role allows.
Organiser access does not authorise ticket payments. A buyer uses a separate supported checkout with their explicit approval, or pays in the browser. Bank, tax and payout details stay in the organiser's own browser.
Recover the original result after an uncertain change. Read the original order or operation before retrying. Reuse its idempotency key when the documented operation supports one. Never repeat an uncertain refund, transfer, message send or ticket grant as a fresh operation.

Change the date or call off an event

Slug change-or-call-off-an-event. Guide: https://www.ticketfairy.com/guides/change-or-call-off-an-event

Choose between postponing, cancelling and refunding, tell ticket holders before they hear it anywhere else, and see what each choice means for their tickets and their money.

For: Choosing between a new date, calling an event off and refunding buyers.

What an agent can do: Fai can tell you the ticket sales, open the page where you make the change, and draft the message to ticket holders. Her figure covers ticket sales only, not add-ons or instalment plans.

Command line:

Show the event, its dates and its status: ticketfairy event get <EVENT_ID>
Postpone the event with the new dates and refund choice in a file you have reviewed (changes the live event): ticketfairy event postpone <EVENT_ID> --from-file postponement.yaml
Cancel the event and confirm at the prompt (changes the live event): ticketfairy event cancel <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

CLI MCP tools that make changes (not available with --read-only): postpone_event, cancel_event.

Prompt for Fai: "For , tell me what its tickets have sold and whether its funds have been paid out yet. Say what your figure leaves out"

Then: Let buyers refund themselves instead

What an agent can do: Fai can open the event so you can switch self-service refunds on, and the Orders page shows how many buyers used them.

Command line:

Show the event and its settings: ticketfairy event get <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Open so I can let buyers refund themselves"

Then: Offer the ticket to another buyer

What an agent can do: Fai can open the event so you can turn resale on, and the dashboard shows the resale figures.

Command line:

Show the event and its settings: ticketfairy event get <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Open so I can let buyers pass their tickets on"

Open your ticket on-sale

Slug open-your-ticket-on-sale. Guide: https://www.ticketfairy.com/guides/open-your-ticket-on-sale

Check every ticket type and buy a test ticket yourself, put the event live at the time you announced, and follow tickets sold, revenue and orders through the first hour.

For: Putting an announced event on sale at a set time.

What an agent can do: Fai can read the event back to you before you open it, publish it once you approve, and report the sales as they come in. The CLI publishes and reads the same event without opening the dashboard.

Command line:

Show the event and its ticket types before you open it: ticketfairy event get <EVENT_ID>
Put the event on sale: ticketfairy event publish <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Before I open for sale, list its ticket types with prices, allocations and sale windows, and tell me anything that looks wrong"

Then: Release more tickets when a tier sells out

What an agent can do: Fai can open the next release once you approve the price and the number, and so can your own AI assistant over the MCP server, after your approval in your MCP client.

Command line:

Show the ticket types so you can pick the one to open: ticketfairy ticket list <EVENT_ID>
Open the held ticket type now. The ticket id comes from the list above: ticketfairy ticket update <EVENT_ID> <TICKET_ID> --sales-start <NOW_ISO_8601>

CLI MCP tools that read: select_event, list_tickets.

CLI MCP tools that make changes (not available with --read-only): update_ticket.

Prompt for Fai: "For , open the next ticket release now and tell me the price and allocation before you change anything"

Then: Collect the people who missed out

What an agent can do: Fai can open the event so you can turn the waiting list on. Today, you switch it on and see who's on it in the dashboard.

Command line:

Show the event and its current settings: ticketfairy event get <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Open so I can turn on its waiting list"

Prepare your door team for event day

Slug prepare-your-door-team. Guide: https://www.ticketfairy.com/guides/prepare-your-door-team

Give everyone on your event door their own scanning login, send the guest list its tickets, set what each ticket may do at the door, test a ticket on every phone, then follow arrivals and find a guest by name once the doors open.

For: Getting a door team ready for the day of an event.

What an agent can do: Fai can invite your door team as Door Staff, send the guest list their tickets, set a ticket's repeat-scan rules, report arrivals and find one guest's ticket. She shows you each invitation, ticket and change and asks you to approve it before anything is sent or saved. You set up wristbands and door times, and undo a test scan, in the dashboard.

Command line:

Read how many have checked in against the tickets sold: ticketfairy checkin status <EVENT_ID>
Read arrivals for each ticket type, such as the guest list against general admission: ticketfairy checkin status <EVENT_ID> --breakdown ticket-types --output json

Prompt for Fai: "Get the door ready for : invite to the event as Door Staff, send guest list tickets to the names I paste below, and let holders leave and come back in, showing me each change before you make it"

Then: Add a name to the guest list on the night

What an agent can do: Fai can send one more guest ticket while you stand at the door. She shows you the name, the ticket type and the count and asks you to approve before she sends it, because you can't unsend a ticket.

Prompt for Fai: "Send a Guest list ticket for to at , show me the ticket type and the count before you send it"

Then: Check whether one guest has arrived

What an agent can do: Fai can tell you the check-in totals and look up one guest's ticket. You undo a scan in the dashboard.

Command line:

Read how many have checked in against the tickets sold: ticketfairy checkin status <EVENT_ID>

Prompt for Fai: "For , how many have checked in so far, and has arrived"

Run last year's event again

Slug run-last-years-event-again. Guide: https://www.ticketfairy.com/guides/run-last-years-event-again

Copy last year's event, change the dates and prices, and open the on-sale to last year's buyers before you announce it publicly.

For: Running an event you have run before.

What an agent can do: Fai can tell you what last year's event sold, tier by tier, and copy it once you approve. The CLI copies the event in one command.

Command line:

Copy last year's event as a draft with its ticket types, team, emails, contacts and FAQs. Enter the start and end as local date and time, such as 2026-10-10T22:00:00; sale windows and door times move by the same number of whole days: ticketfairy event clone <EVENT_ID> --body '{"data": {"displayName": "<NEW EVENT NAME>", "slug": "", "shortName": "", "startDate": "<NEW START>", "endDate": "<NEW END>", "ticketTypes": true, "team": true, "credentials": false, "preRegistration": false, "promotions": false, "delegations": false, "emailTemplates": true, "contacts": true, "faqs": true}}'

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Copy for next year, keep the ticket types and add-ons, and show me the new event in draft with the dates I give you before anything is published"

Then: Repeat an event every week or month

What an agent can do: Fai can open the recurring-event setup for you. You build the series in the dashboard; she creates one event at a time.

Command line:

Copy the event as a draft for one new date, with its ticket types, team, emails, contacts and FAQs. Enter the start and end as local date and time, such as 2026-10-10T22:00:00; sale windows and door times move by the same number of whole days: ticketfairy event clone <EVENT_ID> --body '{"data": {"displayName": "<NEW EVENT NAME>", "slug": "", "shortName": "", "startDate": "<NEW START>", "endDate": "<NEW END>", "ticketTypes": true, "team": true, "credentials": false, "preRegistration": false, "promotions": false, "delegations": false, "emailTemplates": true, "contacts": true, "faqs": true}}'

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Open the recurring-event setup for so I can make it a weekly series"

Then: Save the setup as a template for another promoter

What an agent can do: Fai can open the sharing page for the event.

Command line:

Show the event you are about to share: ticketfairy event get <EVENT_ID>

CLI MCP tools that read: select_event, get_event.

Prompt for Fai: "Open the template sharing page for "

All guides: https://www.ticketfairy.com/guides
Guides MCP server: https://www.ticketfairy.com/docs/mcp (tools search_guides and get_guide)
CLI reference: https://www.npmjs.com/package/ticketfairy
Developer documentation: https://www.ticketfairy.com/developers
Support: support@ticketfairy.com

Guides

Change the date or call off an event

Then: Let buyers refund themselves instead

Then: Offer the ticket to another buyer

Open your ticket on-sale

Then: Release more tickets when a tier sells out

Then: Collect the people who missed out

Prepare your door team for event day

Then: Add a name to the guest list on the night

Then: Check whether one guest has arrived

Run last year's event again

Then: Repeat an event every week or month

Then: Save the setup as a template for another promoter

More

run-event-workflows

Run event workflows

How to work