Google announced at Apple's Worldwide Developers Conference (WWDC) that Apple developers can now integrate Gemini models directly into their applications using the Foundation Models framework. This capability launches with iOS 27, macOS 27, iPadOS 27, visionOS 27, and watchOS 27, utilizing a new LanguageModel protocol that provides a common interface for model inference. By leveraging the Firebase Apple SDK, developers can access cloud-hosted Gemini models through the same API surface used for on-device Apple models, allowing for flexible switching between local and cloud inference to optimize for cost and latency.

The integration utilizes Firebase AI Logic, a production-grade service enabling developers to embed Gemini models into Apple platforms without building a separate backend server. Firebase App Check further secures these service APIs against unauthorized access. Additionally, Google worked with Apple to embed Gemini directly into Xcode, allowing developers to perform complex, multi-step coding tasks—such as code review and bug fixing—without switching between windows or external tools.

Developers can access Gemini via two authentication paths. Individual users can obtain self-serve API keys from Google AI Studio, which provides both a free tier and a paid tier for higher volume needs. Enterprise developers can utilize the Gemini Enterprise Agent Platform to manage dedicated corporate quotas and adhere to specific data privacy requirements. This framework aims to increase development velocity and support the creation of agentic AI experiences (automated systems capable of completing multi-step tasks) across the Apple ecosystem.

Google announced at Apple's Worldwide Developers Conference (WWDC) that Apple developers can now integrate Gemini models directly into their applications using the Foundation Models framework. This capability launches with iOS 27, macOS 27, iPadOS 27, visionOS 27, and watchOS 27, utilizing a new LanguageModel protocol that provides a common interface for model inference. By leveraging the Firebase Apple SDK, developers can access cloud-hosted Gemini models through the same API surface used for on-device Apple models, allowing for flexible switching between local and cloud inference to optimize for cost and latency.

The integration utilizes Firebase AI Logic, a production-grade service enabling developers to embed Gemini models into Apple platforms without building a separate backend server. Firebase App Check further secures these service APIs against unauthorized access. Additionally, Google worked with Apple to embed Gemini directly into Xcode, allowing developers to perform complex, multi-step coding tasks—such as code review and bug fixing—without switching between windows or external tools.

Developers can access Gemini via two authentication paths. Individual users can obtain self-serve API keys from Google AI Studio, which provides both a free tier and a paid tier for higher volume needs. Enterprise developers can utilize the Gemini Enterprise Agent Platform to manage dedicated corporate quotas and adhere to specific data privacy requirements. This framework aims to increase development velocity and support the creation of agentic AI experiences (automated systems capable of completing multi-step tasks) across the Apple ecosystem.