AI assistants that book restaurant tables
Which AI assistants can actually book a restaurant table for me?
Updated · 35 assistants ranked
Thirty-five assistants have a score on our Online task test, which asks for a table for two at a named restaurant this Saturday evening. A pass means a real booking under your name and number, between 6:30 and 8:30pm, with a confirmation sent back; sending you a booking link scores 3. The table below ranks all 35 by that one score.
Top picks
- 110OpenInstinctTop pick
Online tasks 10
- 29
Online tasks 9
- 39DotsFastest that's still good
Online tasks 9Time to finish 6m 51s
- 49OllieBest free option
Online tasks 9
- 58
Online tasks 8
Top pick: highest average here. Best free option: highest among assistants you can use for free. Fastest that's still good: quickest median time to finish among those averaging 7 or more.
Full ranking
| # | Assistant | Online tasks |
|---|---|---|
| 1 | OpenInstinct | 10 |
| 2 | Instinct | 9 |
| 3 | Dots | 9 |
| 4 | Ollie | 9 |
| 5 | Muse | 8 |
| 6 | Hark | 8 |
| 7 | Bantr | 8 |
| 8 | Caddy | 8 |
| 9 | Today | 8 |
| 10 | Neo | 7 |
| 11 | Pally | 7 |
| 12 | szn | 7 |
| 13 | Tomo | 7 |
| 14 | Familiar | 7 |
| 15 | Buoy | 7 |
| 16 | Asmi | 7 |
| 17 | Folk | 7 |
| 18 | Orbits | 7 |
| 19 | OpenClaw | 7 |
| 20 | Luke | 6 |
| 21 | Waldo | 6 |
| 22 | Gemini Spark | 6 |
| 23 | Del | 5 |
| 24 | Persona | 5 |
| 25 | Joshu | 5 |
| 26 | Poke | 4 |
| 27 | Vellum | 4 |
| 28 | Ask Ranger | 4 |
| 29 | Manus | 4 |
| 30 | tinyNature | 4 |
| 31 | Lucas | 4 |
| 32 | AVA | 3 |
| 33 | Wajo / Fo | 2 |
| 34 | Tab | 2 |
| 35 | Brea | 2 |
How this report works
Every assistant with an Online task score is ranked by it. The test uses restaurants that take free-cancellation bookings on Resy or OpenTable, and bookings are cancelled within ten minutes. It is one restaurant booking, not a count of how many restaurants an assistant covers. Scores come from the published score sheet and are dated 2026-10-27. 16 of the 35 ranked assistants carry pre-filled scores, which are placeholder estimates the site has not yet replaced with its own test runs, so treat their places as provisional.
How to choose
- Check that it books rather than links. The test checks for the named restaurant, a table for two, your own name and number, and a confirmation by email or text.
- Give it a time window, not one slot. The test accepts any time from 6:30 to 8:30pm if 7:30 is gone.
- Confirm the restaurant uses a booking system the assistant can reach. Places that only take phone calls need the Phone calls ability instead; see the phone report.
How was a restaurant booking tested?
The assistant is asked to get a table for two at a named restaurant this Saturday at 7:30pm, with anything from 6:30 to 8:30pm acceptable, using the tester's name and number. It passes by making a real booking and reporting the time back with a confirmation.
What if the restaurant only takes phone bookings?
The Online task test uses restaurants on Resy or OpenTable. For restaurants that take only calls, look at the Phone calls test, where an assistant has to call a business and bring back answers.
What users say
People have reported that when they tried to disconnect the agent from their email, they found out it was retaining a copy of their inboxes anyway.
Open sourceIt needed one more code before it got through. Then it reported that the membership was canceled and that there were no scheduled payments, and I received the confirmation email.
Open sourceThe agent, which Meta launched on September 8, booked a Friday date night and made a packing schedule for my upcoming move. It also emailed a friend about a trip next month and made a planning document with places we want to visit.
Open sourcealthough Muse almost could do it, it couldn't complete the final payment section without the site timing out. I ended up having to do it on my computer.
Open sourceFrequently asked questions
Which assistant is best for this?
OpenInstinct leads with an average of 10.0, ahead of Instinct and Dots.
Is there a good free option?
OpenInstinct is the highest-ranked assistant with a free plan, averaging 10.0.
How did you test these assistants?
Every assistant got the same published tests, word for word, scored from three checkpoints and the number of corrections it needed. Each test is on the Dimensions page.