Google announced that developers can now use its new Google Maps tool in the Gemini API to ground their apps in Maps data. They can use the tool to pair Gemini with the latest geospatial data available through Google Maps, which includes over 250 million places.
The search giant already has a similar tool that allows developers to connect the generative AI model to Google Search and access content from the web. This reduces the chances of hallucination as responses are based on real-world data. Moreover, the AI model can access information in real-time and provide citations for the sources it uses.
Google said that developers can pair the two tools to create new AI-powered features for their apps, which are more contextually aware and offer better quality compared to using either tool alone. It added that while Google Search provides descriptive, timely context from the web, Google Maps can pull structured, factual data like addresses, operating hours, and user ratings.
Gemini can automatically determine if a query has a geographical context and use Google Maps data to provide a grounded response. The tool uses the Maps API as a grounding source to generate relevant responses.
The responses also consider other Maps data, such as user ratings, and developers can provide latitude and longitude coordinates of a place to localize search results.
There could be different use cases; for instance, the tool can be used to build conversational trip planners or local guides for travel apps, complete with a full day"s plan, travel time, distance, and local details.
Developers can use the tool for hyperlocal recommendations based on user preferences and designated locations like shops and restaurants. The AI model can answer detailed questions about local places by pulling data from user reviews and other Maps data.
Google announced that Grounding with Google Maps is now generally available, supporting its latest models, including Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash-Lite, and Gemini 2.0 Flash.