Google Launches AI-Powered Offline Dictation App for iOS
On April 7, Google announced the release of its offline-first dictation application, “Google AI Edge Eloquent,” designed to compete with established players like Wispr Flow, SuperWhisper, and Willow. The new app is currently available on iOS, though references to an Android version have been removed from the app store listing, prompting speculation about future availability.
The app is free for download and utilizes Gemma-based automatic speech recognition (ASR) models for dictation. Once these models are installed, users can dictate directly from their devices, with live transcription displayed on-screen. When users opt to pause, the app intelligently filters out filler words such as “um” and “ah,” resulting in polished text that enhances the overall dictation experience.
Users will find an array of transformation options beneath the transcript, including settings for “Key points,” “Formal,” “Short,” and “Long” formats. This versatility allows for tailored output to meet different communication needs, making it suitable for both casual and professional contexts.
Another notable feature is the option to switch to local-only processing by disabling cloud mode. In cloud mode, the app employs cloud-based Gemini models for refining the text further. Additionally, users can import keywords, names, and specific jargon from their Gmail accounts, enabling more personalized dictation, and can even add custom vocabulary to suit their preferences.
The app does not just transcribe; it provides users with a comprehensive history of their transcription sessions. This feature allows users to revisit past dictations and track their words-per-minute speed alongside a total word count. Such functionalities are ideal for professionals who need to monitor their productivity and improve their dictation skills.
According to the App Store description, Google AI Edge Eloquent aims to bridge the gap between natural speech and professional-grade text. The application utilizes advanced AI to interpret user intent, automatically editing out non-essential phrases and distractions, thus producing clean and accurate output.
While the iOS version is currently available, the app’s description hints at an Android version in the works. Incorporating seamless Android integration, the app could potentially become a default keyboard option, allowing users to access it across various text fields, similar to the floating button feature offered by Wispr Flow on Android. As AI-driven transcription technologies continue to engage users, Google’s introduction of this experimental app marks a significant step in advancing the capabilities within this space.