Google Cloud has introduced an exciting new feature, Grounding with Google Search, now available within both Google AI Studio and the Gemini API. This update allows developers to integrate real-time search results into responses generated by Gemini’s AI models, thereby enhancing the accuracy and relevance of the information provided. With this capability, developers can access more up-to-date and factual answers, increasing the reliability of their AI applications.
One of the standout features of Grounding with Google Search is its ability to return not only more precise AI-generated responses but also provide links to the sources that helped ground these answers. This inclusion of supporting search results adds an extra layer of transparency, enabling users to verify the information and further explore the topic. Additionally, search suggestions accompany the responses, helping users to find related content quickly and conveniently.
Grounding with Google Search is available for all generally accessible versions of the Gemini 1.5 models, allowing users to integrate this functionality into their applications seamlessly. In Google AI Studio, users can find the grounding tool under the “Tools” section, while developers can enable it via the Gemini API by activating the google_search_retrieval tool. Testing the feature is free for those using Google AI Studio, while API access comes with a cost of $35 per 1,000 grounded queries for those in the paid tier.
This feature brings several important benefits to developers and users alike. By grounding responses with live data, AI models can provide more factually accurate and real-time information, making them better suited to dynamic environments and more diverse use cases. Moreover, grounding promotes transparency in AI applications, giving users access to the sources of the information and building trust in the generated responses. As a result, applications can offer richer, more reliable information, making them more valuable in a wide range of scenarios.