Resy Restaurant Reservation MCP

Resy Restaurant Reservation MCP

Resy Restaurant Reservation MCP lets your AI assistant find restaurants on Resy, check open tables, and book, list and cancel reservations on your own Resy account. It's a remote MCP server that runs on Apify, so there's nothing to install.

Six tools for one conversation with your assistant: search a city, check a restaurant, book a table, see your reservations, cancel one, and log in. Slots with a prepayment or cancellation fee are never booked until you've seen the amounts and agreed.

Copy to your AI assistant

Copy this block into ChatGPT, Claude, Cursor, or any LLM to start using this actor.

Apify Actor: parsebird/resy-restaurant-reservation-mcp. Remote MCP server for Resy restaurant reservations. Endpoint: https://parsebird--resy-restaurant-reservation-mcp.apify.actor/mcp (Streamable HTTP). Auth header: Authorization: Bearer APIFY_TOKEN. Resy login headers for account tools: X-Resy-Email, X-Resy-Password. Tools: search_restaurants {city (required), cuisine, query, date YYYY-MM-DD, party_size 1-20 default 2, limit 1-25 default 10, sort relevance|rating|slots|name, time HH:MM}; check_availability {restaurant_url (required), date, party_size}; book_reservation {config_token (required), accept_fees boolean}; my_reservations {}; cancel_reservation {resy_token (required)}; login {email, password}. Flow: search or check, let the user choose a slot, book with its config_token, verify with my_reservations. If book_reservation returns status confirmation_required, show the fees to the user and call again with accept_fees true only if they agree. Pricing: pay per search, availability check and confirmed reservation, plus a start event per server run; my_reservations, cancel_reservation and login are free. Docs: https://apify.com/parsebird/resy-restaurant-reservation-mcp. Get your token: https://console.apify.com/settings/integrations.

What is Resy Restaurant Reservation MCP?

Resy Restaurant Reservation MCP is a Resy MCP server and a Resy API alternative for AI agents. Connect Claude, Cursor or another MCP client to it, then ask something like "find Italian restaurants in New York for four on Friday, show me the times, and let me pick one to book."

Your assistant gets live Resy availability with a booking token for every slot. Searching and checking availability don't need a Resy account. Booking creates a real reservation on your account, so you need a Resy account for that.

What can Resy Restaurant Reservation MCP do?

• 🔎 Search a city. Filter Resy restaurants by cuisine, name or keyword, date, party size and time of day, and sort by relevance, rating, open slots or name.
• 🕒 Check one restaurant. See every open table with its deposit and cancellation fee. An empty result says why: sold out, closed, not released yet, or party size not accepted.
• 📅 Book a table. Book any slot from those results with its config_token.
• 💳 Protect you from surprise fees. If a slot has a prepayment or cancellation fee, nothing is booked until your assistant shows you the amounts and you agree.
• 📋 Manage your reservations. List upcoming reservations with their cancellation policies, and cancel them.
• 🔐 Keep your login separate. Each Apify user gets their own server runs, and a login made in one MCP session isn't available to another. Your password and Resy token aren't logged or saved.

Because it runs on Apify, you also get:

• API access with your Apify token
• A Standby server that starts on the first request and stops when idle
• Integrations with Make, Zapier, Slack and webhooks
• Scheduling and export to JSON, CSV, Excel, XML and HTML for the sample run

What data can you get from Resy?

Tool	Inputs	Returns	Cost
search_restaurants	city, optional cuisine, query, date, party_size, limit, sort, time	Restaurants with cuisine, neighborhood, price range, rating, URL and open slots	Search event
check_availability	restaurant_url, optional date, party_size	Status and every open slot with its fees	Availability event
book_reservation	config_token, optional accept_fees	resy_token, reservation_id, restaurant, time, fees and cancellation policy	Reservation event, only when Resy accepts the booking
my_reservations	none	Upcoming reservations with resy_token, time, party size, seating and cancellation policy	Free
cancel_reservation	resy_token	Whether it was cancelled, and any refund Resy reports	Free
login	optional email, password	Whether login worked, your first name and the number of cards on file	Free

Key fields:

Field	Description
slots[].config_token	The token to pass to book_reservation
slots[].time, start, end	Slot time in the restaurant's local time
slots[].type	Seating, such as Dining Room or Chef Counter
slots[].fees	is_paid, deposit_fee and cancellation_fee, from check_availability
slot_count	How many open slots exist. search_restaurants shows up to 10, spread across the day
status	available, sold-out, closed, not-released, not-available or party-size-not-accepted
total_matches	How many restaurants match on Resy, which can be more than limit

How to book a Resy reservation with an AI assistant

1. Get your API token from Apify Console.
2. Add the server to your MCP client, with your Apify token and your Resy login as headers. The examples below show how.
3. Ask your assistant for what you want, such as "a table for two in Miami on Saturday around 7pm."
4. Pick a restaurant and time from what it finds.
5. Tell it to book. If the slot has fees, it shows you the amounts first and books only if you agree.
6. Look for Resy's confirmation email or text. Resy treats a reservation as confirmed once it sends that.

Claude Code

claude mcp add --transport http resy https://parsebird--resy-restaurant-reservation-mcp.apify.actor/mcp \
  --header "Authorization: Bearer <YOUR_APIFY_TOKEN>" \
  --header "X-Resy-Email: <YOUR_RESY_EMAIL>" \
  --header "X-Resy-Password: <YOUR_RESY_PASSWORD>"

Cursor (mcp.json)

{
    "mcpServers": {
        "resy": {
            "url": "https://parsebird--resy-restaurant-reservation-mcp.apify.actor/mcp",
            "headers": {
                "Authorization": "Bearer <YOUR_APIFY_TOKEN>",
                "X-Resy-Email": "<YOUR_RESY_EMAIL>",
                "X-Resy-Password": "<YOUR_RESY_PASSWORD>"
            }
        }
    }
}

Other clients need Streamable HTTP and custom headers. Leave out the Resy headers if you only want to search. You can also use the login tool, but then your password passes through the conversation with the AI model, so headers are the better choice.

Input examples

Search for restaurants:

{ "city": "New York", "cuisine": "Italian", "date": "2026-11-20", "party_size": 4 }

Check one restaurant:

{ "restaurant_url": "https://resy.com/cities/new-york-ny/carbone", "date": "2026-11-20", "party_size": 2 }

Book a slot:

{ "config_token": "rgs://resy/60029/3805357/3/2026-11-20/2026-11-20/11:30:00/2/Dining Room" }

Output examples

search_restaurants, shortened to one restaurant and one slot:

{
    "city": "New York",
    "date": "2026-11-20",
    "party_size": 4,
    "sort": "relevance",
    "total_matches": 303,
    "count": 1,
    "restaurants": [
        {
            "venue_id": 83704,
            "name": "Da Claudio",
            "cuisine": "Italian",
            "neighborhood": "Financial District",
            "city": "New York",
            "price_range": "$$",
            "rating": 4.72,
            "rating_count": 449,
            "url": "https://resy.com/cities/new-york-ny/da-claudio-ny",
            "slot_count": 19,
            "slots": [
                {
                    "start": "2026-11-20 19:00:00",
                    "end": "2026-11-20 20:30:00",
                    "time": "19:00",
                    "type": "Dining Room",
                    "config_token": "rgs://resy/83704/3559914/2/2026-11-20/2026-11-20/19:00:00/4/Dining Room"
                }
            ]
        }
    ]
}

check_availability, shortened to one slot:

{
    "venue_id": 88840,
    "name": "Tsubame",
    "url": "https://resy.com/cities/new-york-ny/tsubame",
    "date": "2026-11-20",
    "party_size": 2,
    "status": "available",
    "slot_count": 2,
    "slots": [
        {
            "start": "2026-11-20 17:30:00",
            "end": "2026-11-20 20:00:00",
            "time": "17:30",
            "type": "Chef Counter",
            "config_token": "rgs://resy/88840/3594091/2/2026-11-20/2026-11-20/17:30:00/2/Dining Room",
            "fees": { "is_paid": true, "deposit_fee": 250.0, "cancellation_fee": null }
        }
    ]
}

book_reservation on that slot, before you agree to the fees. The amounts come from Resy's terms for the slot, and nothing is booked or charged at this step:

{
    "booked": false,
    "status": "confirmation_required",
    "restaurant": "Tsubame",
    "url": "https://resy.com/cities/new-york-ny/tsubame",
    "date": "2026-11-20",
    "time": "17:30",
    "party_size": 2,
    "has_costs": true,
    "total": 653.26,
    "cancellation_fee": 0.0,
    "currency": "USD",
    "cancellation_policy": ["This prepaid reservation is non-refundable. You will not receive a refund if you change or cancel."],
    "message": "This slot has fees, so nothing was booked. Show the amounts and policy to the user, and call book_reservation again with accept_fees true only if they agree."
}

book_reservation after a booking of a slot with no fees. This shows the format: the terms are Resy's for that slot, and resy_token and reservation_id are placeholders:

{
    "booked": true,
    "status": "booked",
    "resy_token": "<TOKEN_FOR_CANCELLING>",
    "reservation_id": 123456789,
    "restaurant": "Le Gratin",
    "url": "https://resy.com/cities/new-york-ny/le-gratin",
    "date": "2026-11-20",
    "time": "11:30",
    "party_size": 2,
    "has_costs": false,
    "total": 0.0,
    "cancellation_fee": 0.0,
    "currency": "USD",
    "cancellation_policy": ["While you won't be charged if you need to cancel, we ask that you do so at least 24 hours in advance."],
    "message": "Resy accepted the booking. Resy treats it as confirmed once it sends its confirmation email or text. Call my_reservations to verify."
}

How to use the Resy MCP API in Python and JavaScript

Python

import httpx

MCP_URL = "https://parsebird--resy-restaurant-reservation-mcp.apify.actor/mcp"
headers = {
    "Authorization": "Bearer <YOUR_APIFY_TOKEN>",
    "Accept": "application/json, text/event-stream",
}

with httpx.Client(headers=headers, timeout=60) as client:
    def rpc(method, params=None):
        response = client.post(MCP_URL, json={"jsonrpc": "2.0", "id": 1, "method": method, "params": params or {}})
        response.raise_for_status()
        return response

    session = rpc("initialize", {
        "protocolVersion": "2025-06-18",
        "capabilities": {},
        "clientInfo": {"name": "my-script", "version": "1.0"},
    })
    client.headers["Mcp-Session-Id"] = session.headers["Mcp-Session-Id"]

    def call_tool(name, arguments):
        result = rpc("tools/call", {"name": name, "arguments": arguments}).json()["result"]
        if result.get("isError"):
            raise RuntimeError(result["content"][0]["text"])
        return result["structuredContent"]

    found = call_tool("search_restaurants", {
        "city": "New York",
        "cuisine": "Italian",
        "date": "2026-11-20",
        "party_size": 4,
        "limit": 5,
    })
    for restaurant in found["restaurants"]:
        print(restaurant["name"], restaurant["rating"], restaurant["slot_count"])

JavaScript

const MCP_URL = 'https://parsebird--resy-restaurant-reservation-mcp.apify.actor/mcp';
const headers = {
    Authorization: 'Bearer <YOUR_APIFY_TOKEN>',
    Accept: 'application/json, text/event-stream',
    'Content-Type': 'application/json',
};

async function rpc(method, params = {}) {
    const response = await fetch(MCP_URL, {
        method: 'POST',
        headers,
        body: JSON.stringify({ jsonrpc: '2.0', id: 1, method, params }),
    });
    if (!response.ok) throw new Error(`HTTP ${response.status}`);
    return response;
}

const session = await rpc('initialize', {
    protocolVersion: '2025-06-18',
    capabilities: {},
    clientInfo: { name: 'my-script', version: '1.0' },
});
headers['Mcp-Session-Id'] = session.headers.get('mcp-session-id');

async function callTool(name, args) {
    const { result } = await (await rpc('tools/call', { name, arguments: args })).json();
    if (result.isError) throw new Error(result.content[0].text);
    return result.structuredContent;
}

const availability = await callTool('check_availability', {
    restaurant_url: 'https://resy.com/cities/new-york-ny/le-gratin',
    date: '2026-11-20',
    party_size: 2,
});
console.log(availability.name, availability.status, availability.slot_count);
availability.slots.slice(0, 3).forEach((slot) => console.log(slot.time, slot.type, slot.config_token));

Add the X-Resy-Email and X-Resy-Password headers to call the account tools from a script. Get your token in Apify Console.

How much does it cost to book a Resy reservation?

Resy Restaurant Reservation MCP uses pay-per-event pricing. Compute and proxy costs are included in the event prices.

Event	Free	Bronze	Silver	Gold
reservation, per booking	$3.49	$2.99	$2.49	$1.99
search, per 1,000	$25.99	$23.99	$21.99	$19.99
availability-check, per 1,000	$25.99	$23.99	$21.99	$19.99
apify-actor-start, per server run	$0.005	$0.005	$0.005	$0.005

• reservation fires once for each booking Resy accepts. Refused bookings, confirmation prompts and repeated calls for the same slot in a session aren't charged.
• search fires for each search that returns at least one restaurant.
• availability-check fires for each completed check, including sold-out and closed answers. Invalid URLs and failed checks aren't charged.
• my_reservations, cancel_reservation and login are free.
• apify-actor-start fires when a server run starts. A run stops after 3 minutes without requests, and the next request starts a new one.

On the Free tier, one search, one availability check and one booking in the same server run cost about $3.55. Any fee a restaurant charges is separate, and is between you and the restaurant.

Is it legal to use an MCP server with Resy?

Using this Actor with your own account is your responsibility, so read Resy's Terms of Service first. As last updated on June 9, 2025, they limit the service to personal, non-commercial use unless Resy agrees in writing. They also prohibit reselling reservations, holding more than one reservation for the same mealtime, booking through multiple accounts, and letting another person use your account, and they restrict automated access to the service. Resy can cancel reservations and suspend accounts that it finds in breach, and it says it deactivates accounts and cancels reservations to fight bots and resellers.

New York's General Business Law § 391-w bars third-party reservation services from listing, promoting or selling a restaurant's reservations without a written agreement with that restaurant, with civil penalties. Other places may have similar rules. Don't use this Actor to resell reservations or to book for other people. This section is not legal advice, so ask a lawyer if you're unsure. You can read more about the general topic in Is web scraping legal?

Other restaurant tools you may like

• Resy Availability Checker: check open tables at up to 10 Resy restaurants in one run, at a low price per restaurant
• OpenTable Restaurants Scraper: cuisine, price, rating, address, phone and photos from OpenTable
• OpenTable Reviews Scraper: ratings, review text and restaurant replies from OpenTable
• TheFork Scraper: restaurant data, menus and reviews from TheFork

FAQ

Will it actually book a real reservation? Yes. book_reservation creates a real reservation on your Resy account. A restaurant's no-show and cancellation rules apply as they do on resy.com, so make sure you can go, or cancel in time.

Do I need a Resy account? Only for book_reservation, my_reservations and cancel_reservation. The server logs in with an email and password, so an account that signs in only with a phone code needs a password first.

Can I save my Resy login instead of sending headers? Not on the main MCP URL, because runs started there don't receive saved input. Apify's docs describe another route: choose Save as a new task, add your login to the task's input, turn on the task's input setting, and connect to the task's own URL. Call login with no arguments to check that it worked. If it says you aren't logged in, use the headers.

Is my Resy password safe? The headers go over HTTPS to your own server runs. The server keeps your login in memory only while the run lasts, and doesn't log or save it. Your password does sit in your MCP client's configuration, so protect that file. Avoid the login tool's email and password arguments if you can, because the AI model sees them.

Why did it ask me to confirm before booking? The slot has a prepayment, deposit or cancellation fee. Some are large and non-refundable, so your assistant has to show you the amounts first and call again with accept_fees: true. search_restaurants doesn't show fees, so a slot from a search can still ask you to confirm. If the fees go up before the booking is made, it asks again.

Which cities does it cover? Cities in Resy's own city list, which had 2,131 entries on November 7, 2026. We tested searches in New York, Los Angeles, Miami, Washington D.C., Dallas, San Francisco, Athens and Tokyo, though Tokyo's results had no open slots. Resy didn't list London or Paris that day, so a search there returns a clear error instead of results from a similar name.

Why does search show only 10 slots when slot_count is higher? Some restaurants list more than 100 slots a day, which would crowd out everything else your assistant reads. Search shows 10 spread across the day. Pass time to focus on a time of day, or use check_availability for all of them.

Why are slots empty when the restaurant is open? No slots can mean sold out, closed or not released yet. check_availability tells you which one. Restaurants release tables on their own schedules.

Can it book restaurants that use a captcha? No. When a restaurant requires a captcha at booking, the result has status browser_required and links to the restaurant, so you can book it on resy.com or in the Resy app. You aren't charged.

What if a booking request gets no clear answer? The server doesn't retry, because that could book twice. It looks for the reservation on your account. If it's there, you get the result as normal. If not, the status is unknown and you aren't charged, so check my_reservations before trying again.

How long does the first request take? The first request starts a server run, which took about 5 seconds in our tests. Later requests in that run are faster.

Can I use it to monitor tables in bulk? It's built for one conversation at a time. For regular monitoring, Resy Availability Checker costs much less per restaurant.

Something doesn't work. Where do I report it? Open a ticket on the Issues tab. Include the tool you called and the message you got, but never your password or tokens.