Google is adding a brand new feature for third-party developers constructing on its Gemini AI models that competitors like OpenAI's ChatGPT, Anthropic's Claude and the growing collection of Chinese open source options are unlikely to get soon: Grounding with Google Maps.
This addition allows developers to mix the reasoning capabilities of Google's Gemini AI models with live geospatial data from Google Maps, allowing applications to supply detailed, location-relevant answers to user queries – akin to business hours, rankings, or the atmosphere of a selected venue.
By leveraging data from over 250 million locations, developers can now create smarter and more responsive location-based experiences.
This is especially useful for applications where proximity, real-time availability, or location-specific personalization are vital – akin to: B. local search, delivery services, real estate and travel planning.
If the user's location is thought, developers can enter latitude and longitude within the request to enhance the standard of the response.
By tightly integrating real-time and historical map data with the Gemini API, Google enables applications to generate informed, location-specific answers with the factual accuracy and contextual depth only possible through its mapping infrastructure.
Merging AI and geointelligence
The recent feature will be accessed in Google AI Studio, where developers can test a live demo based on the Gemini Live API. Models that support grounding with Google Maps include:
-
Gemini 2.5 Pro
-
Gemini 2.5 Flash
-
Gemini 2.5 Flash Lite
-
Gemini 2.0 Flash
In one demonstrationone user asked for recommendations for Italian restaurants in Chicago.
Using map data, the assistant retrieved the top-rated options and clarified a misspelled restaurant name before finding the proper venue with accurate business details.
Developers may obtain a context token to embed a Google Maps widget of their app's UI. This interactive component displays photos, reviews, and other familiar content typically found on Google Maps.
The integration takes place via the generateContent Method in Gemini API where developers include googleMaps as a tool. You may enable a map widget by setting a parameter within the request. The widget, rendered using a returned context token, can provide a visible layer alongside the AI-generated text.
Cross-industry use cases
The Maps grounding tool is designed to support quite a lot of practical use cases:
-
Itinerary creation: Travel apps can create detailed every day plans with route, time and venue information.
-
Personalized local recommendations: Real estate platforms can highlight listings near child-friendly amenities like schools and parks.
-
Detailed location queries: Using community reviews and Maps metadata, applications can provide specific information, akin to whether a restaurant offers outdoor seating.
Developers are really useful to enable the tool only when geographic context is relevant to optimize each performance and value.
According to the developer documentation, pricing starts at $25 per 1,000 prompts – a hefty sum for individuals who make quite a few requests.
Combining search and maps for expanded context
Developers can use grounding with Google Maps together with grounding with Google Search in the identical query.
While the map tool contributes factual data akin to addresses, opening times and reviews, the search tool adds broader context from web content akin to news or event listings.
For example, when asked about live music on Beale Street, the combined tools provide venue details from maps and event times from search.
According to Google, internal tests show that using each tools together results in significantly improved response quality.
Unfortunately, it doesn't appear that Google Maps grounding data includes live vehicle traffic data – at the very least not yet.
Customization and developer flexibility
The experience is customizable. Developers can optimize system prompts, select from different Gemini models, and configure voice settings to customize interactions.
The demo app in Google AI Studio can also be remixable, allowing developers to check ideas, add features, and iterate designs in a versatile development environment.
The API returns structured metadata—including source links, location IDs, and citation ranges—that developers can use to create inline citations or review AI-generated output.
This promotes transparency and increases trust in user-centric applications. Google also requires that Maps-based sources be uniquely attributed and linked to the source via their URI.
Implementation considerations for AI builders
For technical teams integrating this feature, Google recommends:
-
Sharing the user's location context, if known, to supply higher results.
-
Display Google Maps source links directly below relevant content.
-
Only enable the tool if the query clearly involves a geographic context.
-
Monitor latency and disable grounding when performance is critical.
Landing using Google Maps is currently available worldwide, but is banned in several territories (including China, Iran, North Korea and Cuba) and is just not permitted for emergency operations.
Availability and access
Grounding with Google Maps is now generally available via the Gemini API.
With this release, Google further expands the capabilities of the Gemini API, enabling developers to construct AI-driven applications that understand and reply to the world around them.
