Gemini in Apple's Foundation Models framework
Blog post from Firebase
Apple announced a preview integration that lets third-party Gemini cloud models work within its Foundation Models framework on iOS, macOS, iPadOS, watchOS, and visionOS 27 through Firebase AI Logic. Developers can implement Gemini via the shared LanguageModel protocol and replace Apple’s on-device SystemLanguageModel with a GeminiLanguageModel with minimal code changes, while retaining existing SwiftUI interfaces, structured output definitions, and tool configurations. The unified API enables developers to choose local inference for privacy, lower cost, and offline operation or cloud inference for larger context windows and more advanced capabilities, with routing decisions controlled by the app. Firebase App Check can attest requests and block calls from compromised devices, emulators, or scripts, while supported capabilities include web and Google Maps grounding, multimodal prompts using images, audio, video, and PDFs, image generation and editing, streamed and multi-turn conversations, structured Swift outputs, and tool-calling agents. To begin, developers must configure a Firebase project and AI Logic, select a Gemini API provider, install the preview Firebase Apple SDK branch, add the FirebaseAILogic library, configure App Check, and integrate Gemini calls into their Swift applications.
No tracked trend matches for this post yet.
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.