A reservation bot is a program that watches a restaurant's booking page for you and books a table the moment one fits your plans. It doesn't get special access or a secret back door. It sees the same tables you would see in the app. The difference is that it checks constantly, never gets bored, and can finish a booking in well under a second.
That's the whole trick. But the details of how it works, and why some booking platforms barely notice bots while others fight them hard, are more interesting than you might expect.
The problem bots solve
Popular restaurants aren't hard to book because they're full every night forever. They're hard to book because of how tables are released.
Most in-demand restaurants open their books on a rolling schedule: a fixed window ahead, at a fixed time of day. When that moment arrives, thousands of people are refreshing the same page for the same handful of prime-time slots. The good ones disappear in seconds.
Then there's the second wave: cancellations. Plans change, people drop reservations they'd been holding as a backup, and tables quietly reappear at random times of day, often for just a few minutes before someone else takes them. No person can watch for that around the clock. A program can.
What a bot actually does
Strip away the mystery and a reservation bot has four jobs.
1. Ask what's open. Every booking app, whether it's the website or the phone app, gets its information by asking the platform's servers a simple question: for this restaurant, on this date, for this many people, what times are available? The server answers with a list of slots. A bot asks the same question the app does.
2. Ask again, and again. This is called polling. The bot repeats the question on a schedule: frequently around a known release time, less often during the quiet hours. Well-built bots are careful here. They slow down when nothing is changing and focus their effort on the moments when something is likely to open up. Hammering a server is rude, and it's also the fastest way to get blocked.
3. Decide. When slots come back, the bot compares them with what you asked for: which dates, what time window (say 7:00 to 9:00 PM), how many people, maybe whether you'd take a bar seat. This part is mostly simple filtering, and that's the point. The decision has to happen instantly, with no person in the loop.
4. Book. If something matches, the bot goes through the same steps you would: hold the table, confirm it with your account and saved card, and get a confirmation back. On most platforms that's just a couple of quick requests, done in a fraction of a second. A person tapping through the same screens takes 20 or 30 seconds.
Then it tells you, usually by text or email, often before you'd have finished unlocking your phone.
Why speed matters so much
At release time, the window is tiny. When a restaurant opens a new day of tables, the 7:30 and 8:00 slots can be gone within seconds. People lose that race for mechanical reasons: pages loading, fingers tapping, a payment screen that needs a second look. A bot that's already signed in and waiting has none of that delay.
But the bigger, quieter advantage is cancellations. Over a few weeks, a popular restaurant sees a steady trickle of tables come back. Each one is visible only briefly. Someone checking a few times a day will almost never catch them. A bot watching all day catches most of them. For a lot of hard-to-get restaurants, that's where most successful bookings actually come from.
Why platforms treat bots so differently
Here's the part people find most surprising: how much pushback a bot gets depends heavily on which platform the restaurant uses.
Resy has historically been fairly easy to work with. Its availability information is straightforward to request, and in practice it hasn't put much effort into shutting out booking tools. Our read is that this is at least partly a business decision. Resy competes with other platforms to win restaurants, and lots of demand flowing through Resy, from any source, makes it a stronger pitch to restaurant owners. A fully booked dining room is good for everyone involved. That doesn't mean anything goes: requests still need to be reasonable, accounts still need to belong to real diners, and abusive traffic still gets cut off.
DoorDash is a different story. Its reservations live inside a company whose main business is delivery, and that shapes everything. DoorDash has spent years fighting fraud: fake accounts, promo-code abuse, stolen cards. Its booking product inherits all of those defenses. That means far more aggressive bot detection, including checks on what kind of device and browser a request seems to come from, close attention to the network it arrives on, and quick blocks on anything that looks automated. Traffic from ordinary cloud servers is often turned away before it gets a single answer.
For a booking tool, that's the difference between a request that simply works and a large amount of extra engineering just to look like a regular diner on a regular connection. It's a big reason many reservation tools support Resy and very few support DoorDash well.
Are reservation bots fair?
It's a reasonable question. A bot that books a table you'll actually show up to is doing what you'd do yourself if you had unlimited time to refresh an app. The table goes to a real diner either way.
Bots that grab tables to resell them are a different thing entirely. That's the restaurant version of ticket scalping, it hurts restaurants and diners alike, and platforms are right to go after it. The healthy version of this technology books one table, for the person who's going to eat there, and keeps working if plans change so the table can go to someone else.
Getting the most out of one
If you use a reservation bot, a few things make a big difference:
- Be flexible. A wider time window and a few extra dates improve your odds far more than anything else.
- Let it watch for cancellations, not just the release moment. That's where most wins come from.
- Keep your card on file current. Many restaurants require one to hold a table, and an expired card turns a win into a miss.
- Show up, or cancel early. No-show fees are real, and cancelling promptly puts the table back for the next person.
The short version
A reservation bot asks a restaurant's booking system what's available, over and over, and books the moment something fits. Its edge is patience and speed, not secret access. How hard that is to build depends on the platform: some treat extra demand as good for business, while others treat any automation as a threat.
If you'd rather not spend your evenings refreshing an app, Reservation Wrangler does the watching for you.