Google just made a surprise move into the AI dictation space with a quietly released iOS app that works entirely offline. The new app, powered by the company’s lightweight Gemma AI models, puts the search giant in direct competition with rising players like Wispr Flow. According to TechCrunch, the app marks Google’s latest push to bring on-device AI capabilities to mobile users without requiring constant cloud connectivity.

Google dropped a new AI dictation app on the App Store with almost no fanfare, but the implications are anything but quiet. The app leverages Google’s Gemma AI models, the company’s family of lightweight, open-source language models designed specifically for on-device deployment.

What makes this launch particularly interesting is the offline-first approach. While most AI-powered dictation tools rely on cloud processing, Google’s betting that users want their voice data processed locally on their iPhones. It’s a privacy play as much as a technical one, addressing growing concerns about sensitive information being transmitted to remote servers.

The app arrives at a moment when voice AI is heating up considerably. Startups like Wispr Flow have been gaining traction with their own AI-powered dictation solutions, proving there’s real demand for smarter voice input that goes beyond basic speech-to-text. Google’s entrance into this space suggests the tech giant sees the writing on the wall, or rather, the dictation in the air.

Gemma models were specifically designed for this kind of deployment. Google released the Gemma family earlier as a more accessible alternative to its larger Gemini models, optimized to run efficiently on consumer hardware without the computational overhead of cloud-based AI. By bringing Gemma to iOS through a dictation app, Google’s showcasing what on-device AI can actually do in everyday scenarios.

The quiet launch strategy is worth noting. Instead of a splashy product announcement,