TikTok Launches Must Eat, a Restaurant Guide Built Around Video

TikTok is launching Must Eat, a restaurant guide inside its app that puts short videos at the center of dining discovery and positions it against Yelp and Google Maps. The Verge reported the launch.

Must Eat draws on 40 million posts, reviews, and comments about dining that already exist on TikTok. The section has two category tabs, everyday eats and elevated dining. Users can sort and filter results by rating, distance, cuisine type, and whether a business is open at that moment, and they can search for nearby businesses on a map.

Each restaurant page links to TikTok videos about that business, so the creator content that surfaced a restaurant sits beside its listing. Users can also leave reviews, call restaurants, make reservations, and bookmark businesses for later.

Must Eat is initially available in New York City, Los Angeles, and the Bay Area. Restaurants cannot pay to be featured in its recommendation lists.

Must Eat expands TikTok Go, which let users book hotels and attractions through the app. TikTok Newsroom describes TikTok Go as a service for people in the United States to discover and book local services, including hotels, attractions, and tours. Adam Presser is CEO of TikTok US.

TikTok says 27 million US diners visited a restaurant, cafe, or shop for the first time last year thanks to the platform.

The broader context here is a change in where restaurant decisions begin. In the review era, the listing was the destination and the star rating was the main signal. Must Eat reverses that order. The signal is a video, and the listing is where the viewer acts on it. Reservations, calls, and bookmarks all happen without leaving the app, so TikTok is working on the booking step as well as the discovery step. The no-paid-placement rule matters for the same reason. A recommendation system only works if users believe a position was earned, and the rule is the mechanism the company has set up to support that belief. Whether 40 million posts can produce ranking data that holds up across a national dining market, beyond three metro areas, is the question the rollout will answer.

In my view, the most useful test is an ordinary one: whether someone looking for dinner in Los Angeles on a weeknight gets a better answer here than from a map search. The 27 million figure suggests the platform already moves people into restaurants. Must Eat tries to turn that behavior into a product with its own filters, bookmarks, and reservations.

If the feature works as described, the practical gain is simple. A diner can move from a clip to a table in one place, with the video that prompted the visit still attached to the restaurant it describes.