Type "Make an interactive 3D tour of New York landmarks" in Google AI Studio, and a real Google map that can be dragged and turned will be loaded in the preview window.
This is a new capability of AI Studio: applications written in Gemini can directly call Google Maps. The available data is comprehensive, including 3D buildings, street views, location information, routes, and even weather and air quality.
The examples given by Google are store search, self-driving route planning, and a tool for arranging routes for door-to-door pet grooming businesses. These are all things that previously required writing a lot of code and connecting to several APIs.
The threshold is very low this time: one-click configuration, no need to tie up a credit card. The free limit is 100 calls per day per Maps API. If you really want to put it online for others to use, you have to associate a paid Google Cloud account.
100 times is basically prepared for the prototype. A store search page with a little traffic can be used up in one day.
The pain points it solves are very specific. In the past, when letting AI write map applications, the most troublesome part was not the code, but applying for API keys, allocating billing, and processing permissions. Many people got stuck at this step and gave up. Now this layer is eaten up by Google's own ecosystem. Gemini and Maps are originally from the same company, and it is difficult for other companies to copy them.
Independent developers who make gadgets for local life, travel, and stores can use it to build prototypes first.