Google has introduced a new feature for third-party developers utilizing its Gemini AI models. This feature, termed “grounding with Google Maps,” allows developers to integrate live geospatial data from Google Maps with the reasoning abilities of the Gemini AI models. By doing so, applications can generate location-specific responses to user queries, such as business hours and reviews for various venues.
The feature provides access to data from over 250 million locations, enhancing the development of applications that depend on proximity and real-time information, such as delivery services and travel planning. When a user’s location is known, developers can input latitude and longitude coordinates to improve response accuracy. This integration aims to deliver grounded and contextually rich responses using Google’s mapping infrastructure.
Developers can experiment with this feature in Google AI Studio, and it is compatible with several Gemini models, including Gemini 2.5 Pro and Gemini 2.0 Flash. A demonstration showed how the assistant could recommend Italian restaurants in Chicago while correcting misspellings and providing accurate business information. Additionally, developers can embed a Google Maps widget in their applications, which displays photos, reviews, and other FAQs typical of Google Maps.
The grounding tool supports multiple applications, such as itinerary generation for travel apps and personalized recommendations on real estate platforms. Pricing for this feature begins at $25 per 1,000 grounded prompts. Developers are encouraged to use this tool only when geographic context is necessary to optimize performance.
The grounding feature can also work alongside Google Search, allowing developers to combine factual data from Maps with broader web content context for improved response quality. Developers can customize the interaction experience and are advised to monitor various factors such as latency to maintain performance.
Currently, grounding with Google Maps is available globally, although usage is restricted in certain countries. This capability remains a part of Google’s ongoing efforts to enhance the Gemini API, enabling developers to design applications that are responsive to real-world contexts.
Source: https://venturebeat.com/ai/developers-can-now-add-live-google-maps-data-to-gemini-powered-ai-app