Google puts Gemini inside Xcode and Apple's AI framework

If you build apps for Apple platforms, WWDC just made calling cloud AI a whole lot less painful.

Google announced that Apple developers can now securely call cloud-hosted Gemini models directly through Apple’s native Foundation Models framework, spanning iOS 27, macOS 27, iPadOS 27, visionOS 27, and watchOS 27. On top of that, Gemini is stepping inside Xcode to handle agentic coding tasks.

Swapping models in one line of code

At WWDC, Apple opened up its Foundation Models framework to third-party cloud providers via a new public LanguageModel protocol. Google jumped right on it using Firebase AI Logic through the Firebase Apple SDK.

Here is the real win: Apple's on-device models and Google's cloud-hosted Gemini now share the exact same API surface. You can use local inference for speed and privacy, then seamlessly swap to something like Gemini 3.5 Flash when you need heavier reasoning.

Because it runs on Firebase, you do not need to build and maintain a custom backend just to proxy model requests. Abuse protection is handled out of the box via Firebase App Check, and the integration is currently live as a preview release.

Gemini lands in Xcode

Google also teamed up with Apple to embed Gemini straight into Xcode. Instead of flipping back and forth between your editor and a browser tab, developers can onboard Gemini through Xcode’s Intelligence settings panel.

Once configured, Gemini acts as an agentic assistant capable of reviewing code, tracking down bugs, and scaffolding features across multi-step tasks. Google is offering two login tiers: solo devs can grab a self-serve API key with free and paid tiers via Google AI Studio, while companies can route through Gemini Enterprise Agent Platform for guaranteed enterprise quotas and corporate data privacy.

Why it matters

For developers, this removes huge chunks of boring plumbing. You no longer need to roll custom infrastructure just to let an iPhone app talk to a capable frontier model, and having local and cloud models behind the same Swift API makes hybrid architectures practically effortless to test.

Industry-wide, it highlights Apple’s pragmatic side. Apple knows its on-device models cannot handle every demanding multimodal workload, so opening a standardized protocol keeps developers building within Apple frameworks instead of going rogue.

We think Google played this smartly by leaning on Firebase's massive mobile footprint. But the same standardization cuts both ways: because switching to Gemini is literally swapping a model instance in Swift, switching away from Gemini to another cloud provider down the road will be just as easy.