Google quietly launched a new offline-first AI dictation application in early April 2025, offering users the ability to transcribe speech to text without requiring an internet connection. The app, which leverages on-device AI processing, was released with little fanfare through the App Store and Google Play.
The application uses Google's on-device machine learning models to perform real-time speech recognition, meaning audio data does not need to be sent to remote servers. This approach offers both privacy benefits and functionality in areas with limited or no connectivity.
Shortly after launch, Google updated its app store listing to remove references to the Android version of the app, while simultaneously noting that an iOS keyboard extension was coming soon. The move suggested the rollout was still in flux, with the company potentially refining its cross-platform strategy.
The app enters a competitive space that includes Apple's built-in dictation tools, Microsoft's transcription features in Office, and third-party apps such as Otter.ai. Google's offline-first approach, however, differentiates it by prioritizing on-device processing over cloud-based transcription, which remains the industry norm for high-accuracy speech recognition.
