Google has released an offline-first dictation app named ‘Google AI Edge Eloquent’ on iOS, aimed at competing with industry counterparts like Wispr Flow and SuperWhisper. The app leverages Gemma-based automatic speech recognition (ASR) models, enabling users to initiate dictation once the models are downloaded. Users can witness live transcription, with the app intelligently filtering out filler words like ‘um’ and ‘ah’ when paused, ensuring polished text output.
The app offers transformation options such as ‘Key points’, ‘Formal’, ‘Short’, and ‘Long’ for the transcribed text. Users have the flexibility to disable cloud mode for local-only processing, while also being able to import personalized keywords, names, and jargon from their Gmail accounts. Additionally, users can supplement the app’s vocabulary with custom words.
Google AI Edge Eloquent maintains a transcription history, facilitating easy search and access to past sessions. The app provides insights like words transcribed in the last session, word-per-minute speed, and total words spoken. The app enhances user experience by capturing intended meaning and automatically refining text by eliminating speech fillers and self-corrections.
Although currently exclusive to iOS, the App Store description hints at an upcoming Android version. The launch of ‘Google AI Edge Eloquent’ introduces a new option for speech-to-text technology.
Source: TechCrunch