Use SupaPark in Cursor
Add SupaPark to Cursor with one click and Cursor's agent can call SupaPark's Walt Disney World tools: live waits, crowd forecasts, Lightning Lane prices and sellouts, dining reservation odds, and your own watches and trip. Useful for planning in the editor you already have open, or for building something on top of real park data.
Part of Gold, for Walt Disney World. Connect once and revoke any time in Settings.
Set it up in Cursor
- Step 1: Be on SupaPark Gold. Gifted Gold and Family Sharing Gold count too.
- Step 2: Click Add to Cursor, confirm the install, then select Connect next to SupaPark to sign in.
https://mcp.supapark.com/mcp- The Add to Cursor button installs the server; then select Connect next to SupaPark in Cursor's MCP settings to sign in.
- For a project-only setup, put the config in .cursor/mcp.json instead of ~/.cursor/mcp.json.
Prefer a personal API key?
Create a key in SupaPark Settings → Connect to AI assistants → Advanced, then add to ~/.cursor/mcp.json (or .cursor/mcp.json in a project):
{
"mcpServers": {
"supapark": {
"url": "https://mcp.supapark.com/mcp",
"headers": {
"Authorization": "Bearer spk_live_YOUR_KEY"
}
}
}
}Try asking
- “What are the waits at Magic Kingdom right now? Anything short that's worth it?”
- “When should we ride Flight of Passage tomorrow, early or late?”
- “Which Disney World restaurants are hardest to book right now?”
- “What does Lightning Lane Multi Pass cost for Animal Kingdom on Saturday?”
- “What's the height requirement for Expedition Everest, and does it have single rider?”
- “Did my Be Our Guest watch find anything?”
See it answer
Examples built from real SupaPark tool output, captured at the time shown. They are not live: ask today and your assistant gets the numbers SupaPark has at that moment. The reply wording is an example; the numbers are what the tool returned.
get_hardest_dining_reservationsSee the tool output (excerpt)
{
"as_of_label": "Nov 4, 1:03 PM ET",
"quality_state": "ok",
"items": [
{
"rank": 1,
"restaurant": "Nomad Lounge",
"park_name": "Animal Kingdom",
"tier": "Very Hard",
"open_rate": "<1%",
"drops_30d": 179
},
{
"rank": 2,
"restaurant": "Oga's Cantina",
"park_name": "Hollywood Studios",
"tier": "Very Hard",
"open_rate": "<1%",
"drops_30d": 24
},
{
"rank": 3,
"restaurant": "Cinderella's Royal Table",
"park_name": "Magic Kingdom",
"tier": "Very Hard",
"open_rate": "1%",
"drops_30d": 2470
},
{
"rank": 4,
"restaurant": "Space 220 Restaurant",
"park_name": "EPCOT",
"tier": "Very Hard",
"open_rate": "17%",
"drops_30d": 2040
}
],
"notes": [
"Open rate = share of SupaPark's reservation-availability checks that found a table. Drops = cancellations SupaPark caught in the last 30 days."
]
}get_live_wait_timesSee the tool output (excerpt)
{
"citation": "Source: SupaPark — https://supapark.com/rides/hollywood-studios (as of Nov 4, 1:03 PM ET).",
"data": [
{
"attractionName": "Star Wars: Rise of the Resistance",
"posted": 70,
"realWait": 56,
"realWaitLow": 48,
"realWaitHigh": 64,
"confidence": "high",
"ageMinutes": 3
},
{
"attractionName": "Slinky Dog Dash",
"posted": 55,
"realWait": 40,
"realWaitLow": 34,
"realWaitHigh": 46,
"confidence": "high",
"ageMinutes": 3
},
{
"attractionName": "Mickey & Minnie's Runaway Railway",
"posted": 50,
"realWait": 39,
"realWaitLow": 33,
"realWaitHigh": 44,
"confidence": "high",
"ageMinutes": 3
},
{
"attractionName": "Toy Story Mania!",
"posted": 10,
"realWait": 8,
"realWaitLow": 6,
"realWaitHigh": 11,
"confidence": "high",
"ageMinutes": 3
},
{
"attractionName": "Rock ’n’ Roller Coaster Starring The Muppets",
"posted": 45,
"realWait": null,
"confidence": "low",
"ageMinutes": 3
}
],
"meta": {
"park": "hs",
"count": 24,
"note": "real_wait is an ESTIMATE from historical posted-vs-actual patterns, shown only when confident."
}
}Questions
Does it work with any model I pick in Cursor?
The SupaPark tools are available to whichever model Cursor runs with MCP tools enabled. The numbers come from SupaPark either way; how the answer reads depends on the model.
What are the rate limits?
Each signed-in app or key gets 60 requests a minute and 5,000 a day.
What if the data is not available?
The tool tells your assistant the data is unavailable and why, and every live answer carries the time the data was read. It never fills the gap with an estimate presented as live data.
What you can ask
By task
Everything in one place: SupaPark for AI assistants. Every tool, scope and error: MCP docs.
App names belong to their owners and are listed only to show compatibility.